TopicSuggestions.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "from owlready2 import *\n",
    "import re\n",
    "import fasttext \n",
    "import numpy as np\n",
    "from numpy import dot\n",
    "from numpy.linalg import norm\n",
    "import networkx as nx\n",
    "import hnswlib\n",
    "from os import path\n",
    "from gensim.models.poincare import PoincareModel"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sentence_transformers import SentenceTransformer\n",
    "model = SentenceTransformer('roberta-large-nli-stsb-mean-tokens')\n",
    "def embed(inputTexts):\n",
    "    return model.encode(inputTexts)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "def cosineSimilarity(a, b):\n",
    "    return np.inner(a, b)/(norm(a)*norm(b))\n",
    "\n",
    "\n",
    "# Usage: split(name.lower) for name in class_names_list\n",
    "def delimSplit(string, delimiters = ['-','_'], maxsplit=0):\n",
    "    import re\n",
    "    regexPattern = '|'.join(map(re.escape, delimiters))\n",
    "    return [word for word in re.split(regexPattern, string, maxsplit) if len(word) > 0]\n",
    "\n",
    "def getName(string):\n",
    "    nameList= delimSplit(string)\n",
    "    return ' '.join(nameList)\n",
    "\n",
    "def preprocessString(string):\n",
    "    string = string.lower()\n",
    "    string = re.sub(\"[^a-zA-Z]\", \" \", string)            # remove non alphabets with a space which is used as a delimiter later\n",
    "    return string"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Nonde\n"
     ]
    }
   ],
   "source": [
    "class CSStudent: \n",
    "  \n",
    "    ontologyFileName = None\n",
    "    def printHello(self, string):\n",
    "        print(string)\n",
    "    # The init method or constructor \n",
    "    def __init__(self): \n",
    "        if self.ontologyFileName is None:\n",
    "            self.printHello('Nonde')\n",
    "tmp = CSStudent()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "class TopicRecommender: \n",
    "  \n",
    "    indexer = None\n",
    "    ontologyFileName = 'ontology.owl'\n",
    "    \n",
    "    def buildIndexer(self, docEmbeddings):\n",
    "        numElements = len(docEmbeddings)\n",
    "        dimension = len(docEmbeddings[0])\n",
    "        embeddings = np.asarray(docEmbeddings)\n",
    "        data_labels = np.arange(numElements)\n",
    "        p = hnswlib.Index(space = 'cosine', dim = dimension) # possible options are l2, cosine or ip\n",
    "        p.init_index(max_elements = numElements, ef_construction = 200, M = dimension)\n",
    "        p.add_items(embeddings, data_labels)\n",
    "        p.set_ef(30) \n",
    "        return p\n",
    "\n",
    "    def saveIndexer(self, p):\n",
    "        index_path='./models/topicsHnwslib.bin'\n",
    "        print(\"Saving index to '%s'\" % index_path)\n",
    "        p.save_index(\"./models/topicsHnwslib.bin\")\n",
    "\n",
    "    def loadIndexer(self, filepath, numElements):\n",
    "        p = hnswlib.Index(space='cosine', dim=1024)  # the space can be changed - keeps the data, alters the distance function.\n",
    "        p.load_index(\"./models/topicsHnwslib.bin\", max_elements = numElements)\n",
    "        return p\n",
    "    def getQueryEmbedding(self, query, entity = False):\n",
    "        if entity:\n",
    "            raise NotImplementedError\n",
    "        else:\n",
    "            return embed([query])[0]\n",
    "    \n",
    "    def getSeedTopics(self, query, numSeeds = 2):\n",
    "        queryEmbedding = self.getQueryEmbedding(query, entity = False)\n",
    "        labels, _ = self.indexer.knn_query(queryEmbedding, k = numSeeds )\n",
    "        labels = labels[0]     ## only one doc embedding searched for, not a list\n",
    "\n",
    "        return [self.topicNames[ind] for ind in labels]\n",
    "\n",
    "    def expandTopicSet(self, topicList, K = 5):\n",
    "        poincareEmbeddings = []\n",
    "        for topic in topicList:\n",
    "            poincareEmbeddings.append(self.model.kv.get_vector(topic))\n",
    "        averagedEmbedding = np.mean(np.asarray(poincareEmbeddings), axis = 0)\n",
    "        suggestedTopics = self.model.kv.most_similar(averagedEmbedding, topn=K)\n",
    "        suggestedTopics = [tup[0] for tup in suggestedTopics]\n",
    "        return suggestedTopics\n",
    "\n",
    "\n",
    "    def suggestions(self, query, K=8):\n",
    "        seedTopics =  self.getSeedTopics(query, numSeeds = 2)\n",
    "        expandedTopics = self.expandTopicSet(seedTopics, K = K - 2)\n",
    "        topicSet = set(seedTopics)\n",
    "        topicList = seedTopics\n",
    "        for topic in expandedTopics:\n",
    "            if topic not in topicSet:\n",
    "                topicList.append(topic)\n",
    "        return topicList\n",
    "     \n",
    "    def __init__(self, ontologyFileName = 'ontology.owl'): \n",
    "        self.ontologyFileName = ontologyFileName\n",
    "        file_prefix = 'file://'\n",
    "        self.onto = get_ontology(file_prefix + \"./data/\" + self.ontologyFileName ).load()\n",
    "        class_list = list(self.onto.classes())\n",
    "        adjList = dict()\n",
    "        for c in class_list:\n",
    "            adjList[getName(c.name.lower())] = [getName(s.name.lower()) for s in c.subclasses()]\n",
    "        G = nx.from_dict_of_lists(adjList)                     ## vertex names are of the form: absolute error loss\n",
    "        phraseToClassDict = {}                             ## 'absolute error loss' mapped to ontology.Absolute_Error_Loss\n",
    "        for c in class_list:\n",
    "            phraseToClassDict[getName(c.name.lower())] = c\n",
    "        self.topicNames = list(phraseToClassDict.keys())\n",
    "\n",
    "        relations = list(G.edges())\n",
    "        self.model = PoincareModel(relations, negative=2)\n",
    "        self.model.train(epochs=50)\n",
    "        \n",
    "        if self.indexer is None:\n",
    "            if path.exists(\"./models/topicsHnwslib.bin\"):\n",
    "                self.indexer = self.loadIndexer(filepath = \"./models/topicsHnwslib.bin\", numElements= len(self.topicNames))\n",
    "            else:\n",
    "                topicEmbeddings = embed(self.topicNames)\n",
    "                self.indexer = self.buildIndexer(topicEmbeddings)\n",
    "                self.saveIndexer(self.indexer)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "tr = TopicRecommender()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['chatbots',\n",
       " 'conversational bot',\n",
       " 'personal assistant',\n",
       " 'verbal assistant',\n",
       " 'bots',\n",
       " 'collaborative filtering']"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tr.suggestions('chatbot') ## Keep 3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['heuristic search',\n",
       " 'pure heuristic search',\n",
       " 'iterative deepening a* search',\n",
       " 'greedy search',\n",
       " 'best first search',\n",
       " 'a* search']"
      ]
     },
     "execution_count": 60,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tr.suggestions('heuristic search')  ## Keep 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['cnn', 'rcnn', 'mask rcnn', 'googlenet/inception', 'resnet', 'zfnet', 'lenet']"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tr.suggestions('cnn')      ## Keep 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['genes prediction',\n",
       " 'neuroscience',\n",
       " 'genetics',\n",
       " 'fluorescence imaging',\n",
       " 'microbiology']"
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tr.expandTopicSet(['genetics', 'neuroscience'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "def buildIndexer(docEmbeddings):\n",
    "    numElements = len(docEmbeddings)\n",
    "    dimension = len(docEmbeddings[0])\n",
    "    embeddings = np.asarray(docEmbeddings)\n",
    "    data_labels = np.arange(numElements)\n",
    "    p = hnswlib.Index(space = 'cosine', dim = dimension) # possible options are l2, cosine or ip\n",
    "    p.init_index(max_elements = numElements, ef_construction = 200, M = dimension)\n",
    "    p.add_items(embeddings, data_labels)\n",
    "    p.set_ef(30) \n",
    "    return p\n",
    "\n",
    "def saveIndexer(p):\n",
    "    index_path='./models/topicsHnwslib.bin'\n",
    "    print(\"Saving index to '%s'\" % index_path)\n",
    "    p.save_index(\"./models/topicsHnwslib.bin\")\n",
    "\n",
    "def loadIndexer(filepath, numElements):\n",
    "    p = hnswlib.Index(space='cosine', dim=1024)  # the space can be changed - keeps the data, alters the distance function.\n",
    "    p.load_index(\"./models/topicsHnwslib.bin\", max_elements = numElements)\n",
    "    return p\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "def getQueryEmbedding(query, entity = False):\n",
    "    if entity:\n",
    "        raise NotImplementedError\n",
    "    else:\n",
    "        return embed([query])[0]\n",
    "    \n",
    "def getSeedTopics(query, topicNames, indexer, numSeeds = 2):\n",
    "    queryEmbedding = getQueryEmbedding(query, entity = False)\n",
    "    if indexer is None:\n",
    "        if path.exists(\"./models/topicsHnwslib.bin\"):\n",
    "            indexer = loadIndexer(filepath = \"./models/topicsHnwslib.bin\", numElements= len(topicNames))\n",
    "        else:\n",
    "            topicEmbeddings = embed(topicNames)\n",
    "            indexer = buildIndexer(topicEmbeddings)\n",
    "            saveIndexer(indexer)\n",
    "    labels, _ = indexer.knn_query(queryEmbedding, k = numSeeds )\n",
    "    labels = labels[0]     ## only one doc embedding searched for, not a list\n",
    "        \n",
    "    return [topicNames[ind] for ind in labels]\n",
    "\n",
    "def getClosestTopics(model, topicList, K = 5):\n",
    "    poincareEmbeddings = []\n",
    "    for topic in topicList:\n",
    "        poincareEmbeddings.append(model.kv.get_vector(topic))\n",
    "    averagedEmbedding = np.mean(np.asarray(poincareEmbeddings), axis = 0)\n",
    "    suggestedTopics = model.kv.most_similar(averagedEmbedding, topn=K)\n",
    "    suggestedTopics = [tup[0] for tup in suggestedTopics]\n",
    "    return suggestedTopics\n",
    "\n",
    "\n",
    "def suggestions(query, K=5):\n",
    "    seedTopics =  getSeedTopics(query, topicNames, indexer, numSeeds = 2)\n",
    "    expandedTopics = getClosestTopics(model, topicList, K = K - 2)\n",
    "    return seedTopics + expandedTopics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "def chunkedEmbeddings(enityList):   \n",
    "    n = 100     # block size\n",
    "    entities = [entityList[i:i + n] for i in range(0, len(entityList), n)]\n",
    "    embeddingList = []\n",
    "    count = 0\n",
    "    for entitySubList in tqdm(entities):\n",
    "        embeddings = embed(entitySubList)#.numpy().tolist()\n",
    "        for embedding in embeddings:\n",
    "            embeddingList.append(embedding)\n",
    "            count += 1\n",
    "    assert count == len(entityList)\n",
    "    return embeddingList                  \n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [],
   "source": [
    "file_prefix = 'file://'\n",
    "onto = get_ontology(file_prefix + \"./data/ontology.owl\").load()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "class_list = list(onto.classes())\n",
    "\n",
    "phraseToClassDict = {}                             ## 'absolute error loss' mapped to ontology.Absolute_Error_Loss\n",
    "for c in class_list:\n",
    "    phraseToClassDict[getName(c.name.lower())] = c"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of nodes:  1074 . Number of edges:  1090\n"
     ]
    }
   ],
   "source": [
    "adjList = dict()\n",
    "for c in class_list:\n",
    "    adjList[getName(c.name.lower())] = [getName(s.name.lower()) for s in c.subclasses()]\n",
    "G = nx.from_dict_of_lists(adjList)                     ## vertex names are of the form: absolute error loss\n",
    "print('Number of nodes: ', len(list(G.nodes)), '. Number of edges: ', len(list(G.edges())))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "topicNames = list(phraseToClassDict.keys())\n",
    "topicEmbeddings = embed(topicNames)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['logistic regression', 'logistic function']"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "getSeedTopics('logistic regression', topicNames, None, numSeeds = 2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "path.exists(\"./models/topicsHnwslib.bin\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "from gensim.models.poincare import PoincareModel\n",
    "relations = list(G.edges())\n",
    "model = PoincareModel(relations, negative=2)\n",
    "model.train(epochs=50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "ename": "ValueError",
     "evalue": "Can only plot 2-D vectors",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mValueError\u001b[0m                                Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-23-944a3e088a31>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0mgensim\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mviz\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpoincare\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mpoincare_2d_visualization\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 2\u001b[0;31m \u001b[0mpoincare_2d_visualization\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmodel\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mset\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mrelations\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'title'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnum_nodes\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m50\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mshow_node_labels\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m~/anaconda3/lib/python3.7/site-packages/gensim/viz/poincare.py\u001b[0m in \u001b[0;36mpoincare_2d_visualization\u001b[0;34m(model, tree, figure_title, num_nodes, show_node_labels)\u001b[0m\n\u001b[1;32m     50\u001b[0m     \u001b[0mvectors\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mkv\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msyn0\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     51\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mvectors\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m!=\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 52\u001b[0;31m         \u001b[0;32mraise\u001b[0m \u001b[0mValueError\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'Can only plot 2-D vectors'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     53\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     54\u001b[0m     \u001b[0mnode_labels\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mkv\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mindex2word\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mValueError\u001b[0m: Can only plot 2-D vectors"
     ]
    },
    {
     "ename": "ValueError",
     "evalue": "Can only plot 2-D vectors",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mValueError\u001b[0m                                Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-23-944a3e088a31>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0mgensim\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mviz\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpoincare\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mpoincare_2d_visualization\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 2\u001b[0;31m \u001b[0mpoincare_2d_visualization\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmodel\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mset\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mrelations\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'title'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnum_nodes\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m50\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mshow_node_labels\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m~/anaconda3/lib/python3.7/site-packages/gensim/viz/poincare.py\u001b[0m in \u001b[0;36mpoincare_2d_visualization\u001b[0;34m(model, tree, figure_title, num_nodes, show_node_labels)\u001b[0m\n\u001b[1;32m     50\u001b[0m     \u001b[0mvectors\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mkv\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msyn0\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     51\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mvectors\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m!=\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 52\u001b[0;31m         \u001b[0;32mraise\u001b[0m \u001b[0mValueError\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'Can only plot 2-D vectors'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     53\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     54\u001b[0m     \u001b[0mnode_labels\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mkv\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mindex2word\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mValueError\u001b[0m: Can only plot 2-D vectors"
     ]
    }
   ],
   "source": [
    "from gensim.viz.poincare import poincare_2d_visualization\n",
    "poincare_2d_visualization(model, set(relations), 'title', num_nodes=50, show_node_labels=())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.manifold import TSNE\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "def tsne_plot(model):  \n",
    "    \"Creates and TSNE model and plots it\"\n",
    "    labels = []\n",
    "    tokens = []\n",
    "\n",
    "    for word in model.kv.vocab:\n",
    "        tokens.append(model.kv.get_vector(word))\n",
    "        labels.append(word)\n",
    "    \n",
    "    tsne_model = TSNE(perplexity=40, n_components=2, init='pca', n_iter=2500, random_state=23)\n",
    "    new_values = tsne_model.fit_transform(tokens)\n",
    "\n",
    "    x = []\n",
    "    y = []\n",
    "    for value in new_values:\n",
    "        x.append(value[0])\n",
    "        y.append(value[1])\n",
    "        \n",
    "    plt.figure(figsize=(16, 16)) \n",
    "    for i in range(len(x)):\n",
    "        plt.scatter(x[i],y[i])\n",
    "        plt.annotate(labels[i],\n",
    "                     xy=(x[i], y[i]),\n",
    "                     xytext=(5, 2),\n",
    "                     textcoords='offset points',\n",
    "                     ha='right',\n",
    "                     va='bottom')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['conversational bot']"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "adjList['chatbots']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('conversational bot', 0.14262441351319782),\n",
       " ('personal assistant', 0.22780726713794536),\n",
       " ('verbal assistant', 0.2642991692430936),\n",
       " ('bots', 0.399117495650599),\n",
       " ('collaborative filtering', 0.6891708398831656),\n",
       " ('session based recommendations', 0.7130494416753574),\n",
       " ('content based filtering', 0.7230437834678549),\n",
       " ('video object segmentation', 1.3233252389603947),\n",
       " ('sociology', 1.3470926795690297),\n",
       " ('unsupervised video object segmentation', 1.3474609584183526)]"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "[('conversational bot', 0.14262441351319782),\n",
       " ('personal assistant', 0.22780726713794536),\n",
       " ('verbal assistant', 0.2642991692430936),\n",
       " ('bots', 0.399117495650599),\n",
       " ('collaborative filtering', 0.6891708398831656),\n",
       " ('session based recommendations', 0.7130494416753574),\n",
       " ('content based filtering', 0.7230437834678549),\n",
       " ('video object segmentation', 1.3233252389603947),\n",
       " ('sociology', 1.3470926795690297),\n",
       " ('unsupervised video object segmentation', 1.3474609584183526)]"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.kv.most_similar('chatbots')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'absolute error loss': ['mean absolute error (mae)'],\n",
       " 'activation functions': ['arsinh function',\n",
       "  'arctan function',\n",
       "  'bent identity function',\n",
       "  'binary step function',\n",
       "  'bipolar step function',\n",
       "  'elliotsig function',\n",
       "  'gaussian function',\n",
       "  'identity function',\n",
       "  'logistic function',\n",
       "  'maxout function',\n",
       "  'ramp function',\n",
       "  'sq rbf function',\n",
       "  'sqnl function',\n",
       "  'sinc function',\n",
       "  'sinusoid function',\n",
       "  'softexponential function',\n",
       "  'softplus function',\n",
       "  'soft clipping function',\n",
       "  'softmax function',\n",
       "  'tanh function',\n",
       "  'adaptive piecewise linear (apl)',\n",
       "  'bipolar rectified linear unit (brelu)',\n",
       "  'exponential linear unit (elu)',\n",
       "  'gaussian error linear units (gelu)',\n",
       "  'inverse square root linear unit (isrlu)',\n",
       "  'inverse square root unit (isru)',\n",
       "  'leaky rectified linear unit (leaky relu)',\n",
       "  'parameteric rectified linear unit (prelu)',\n",
       "  'randomized leaky rectified linear unit (rrelu)',\n",
       "  'rectified linear unit (relu)',\n",
       "  's shaped rectified linear activation unit (srelu)',\n",
       "  'scaled exponential linear unit (selu)',\n",
       "  'sigmoid linear unit (silu)'],\n",
       " 'adadelta': [],\n",
       " 'adagrad': [],\n",
       " 'adam': [],\n",
       " 'agriculture': [],\n",
       " 'ant colony optimization': [],\n",
       " 'arsinh function': [],\n",
       " 'arctan function': [],\n",
       " 'architecture': [],\n",
       " 'artificial bee colony optimization': [],\n",
       " 'artificial fish swarm algorithm': [],\n",
       " 'artificial intelligence': ['applications', 'fundamental', 'thematics', 'wp'],\n",
       " 'artificial neural networks': ['boltzmann machine',\n",
       "  'hebbian learning',\n",
       "  'hopfield network',\n",
       "  'multilayer feed forward networks',\n",
       "  'perceptrons',\n",
       "  'rnn'],\n",
       " 'arts and entertainment': ['entertainment', 'arts'],\n",
       " 'assistance': ['bots', 'recommendation systems'],\n",
       " 'astronomy': [],\n",
       " 'automobiles': [],\n",
       " 'autonomous vehicles': ['car racing',\n",
       "  'continuous control',\n",
       "  'pedestrian detection',\n",
       "  'steering control'],\n",
       " 'backpropogation': ['levenberg marquardt backpropagation',\n",
       "  'resilient backpropagation (rprop)'],\n",
       " 'bacterial foraging optimization': [],\n",
       " 'banking': [],\n",
       " 'behavior based ai': ['subsumption architecture'],\n",
       " 'bent identity function': [],\n",
       " 'best first search': [],\n",
       " 'bidirectional search': [],\n",
       " 'big data': [],\n",
       " 'binary classification loss functions': ['hinge loss',\n",
       "  'binary cross entropy (bce)'],\n",
       " 'binary sigmoidal function': [],\n",
       " 'binary step function': [],\n",
       " 'bioinformatics': ['dna sequencing', 'genes prediction'],\n",
       " 'biology': ['bioinformatics',\n",
       "  'cell biology',\n",
       "  'ecology',\n",
       "  'fluorescence imaging',\n",
       "  'genetics',\n",
       "  'high throughput imaging',\n",
       "  'microbiology',\n",
       "  'molecular biology',\n",
       "  'neuroscience'],\n",
       " 'bipolar sigmoidal function': [],\n",
       " 'bipolar step function': [],\n",
       " 'blind search': ['brute force search'],\n",
       " 'bots': ['chatbots', 'personal assistant', 'verbal assistant'],\n",
       " 'breadth first search': [],\n",
       " 'brute force search': ['bidirectional search',\n",
       "  'breadth first search',\n",
       "  'depth first search',\n",
       "  'iterative deepening depth first search',\n",
       "  'uniform cost search'],\n",
       " 'business': [],\n",
       " 'cancer': [],\n",
       " 'cell imaging': [],\n",
       " 'cell painting': [],\n",
       " 'cell biology': ['cell imaging', 'cell painting'],\n",
       " 'census': [],\n",
       " 'chatbots': ['conversational bot'],\n",
       " 'classification loss functions': ['binary classification loss functions',\n",
       "  'margin classifier',\n",
       "  'multi class classification loss functions',\n",
       "  'negative log likelihood',\n",
       "  'soft margin classifier'],\n",
       " 'climate': ['weather'],\n",
       " 'computational creativity': [],\n",
       " 'computer security': [],\n",
       " 'construction': [],\n",
       " 'conversational bot': [],\n",
       " 'cosine error': [],\n",
       " 'crime': [],\n",
       " 'cross entropy loss': [],\n",
       " 'culture and humanities': ['culture'],\n",
       " 'cyber security': ['computer security'],\n",
       " 'dna sequencing': [],\n",
       " 'data analytics': [],\n",
       " 'data science': ['data mining'],\n",
       " 'data mining': [],\n",
       " 'demographics': [],\n",
       " 'dempster shafer theory': [],\n",
       " 'depth first search': [],\n",
       " 'design and modelling': [],\n",
       " 'diabetes': [],\n",
       " 'digital preservation': [],\n",
       " 'digital media': [],\n",
       " 'disaster management': ['disaster response',\n",
       "  'forecast system',\n",
       "  'warning systems'],\n",
       " 'distributed artificial intelligence': [],\n",
       " 'ecology': [],\n",
       " 'economics': ['revenue forecast'],\n",
       " 'education': [],\n",
       " 'elliotsig function': ['softsign function'],\n",
       " 'embedding loss functions': ['cosine error', 'l1 hinge error'],\n",
       " 'entertainment': ['movies', 'music', 'photography', 'games'],\n",
       " 'entropy based thresholding': [],\n",
       " 'epidemiology': [],\n",
       " 'evolutionary computation': [],\n",
       " 'feature engineering': [],\n",
       " 'film': [],\n",
       " 'finance': [],\n",
       " 'first order optimization algorithms': ['adadelta',\n",
       "  'adagrad',\n",
       "  'adam',\n",
       "  'gradient descent',\n",
       "  'mini batch gradient descent',\n",
       "  'nesterov accelerated gradient',\n",
       "  'stochastic gradient descent'],\n",
       " 'flower pollination optimization': [],\n",
       " 'fluorescence imaging': [],\n",
       " 'fuzzy logic': [],\n",
       " 'fuzzy sets': ['fuzzy logic'],\n",
       " 'fuzzy systems': [],\n",
       " 'gru': [],\n",
       " 'gaussian function': [],\n",
       " 'genetic algorithm optimization': [],\n",
       " 'genetics': ['genomics'],\n",
       " 'genomics': [],\n",
       " 'geography': [],\n",
       " 'geospatial': [],\n",
       " 'global valley approach': [],\n",
       " 'gradient descent': [],\n",
       " 'gravitational search algorithm': [],\n",
       " 'greedy search': ['best first search'],\n",
       " 'health': ['health care'],\n",
       " 'heuristic search': ['greedy search',\n",
       "  'pure heuristic search',\n",
       "  'a* search',\n",
       "  'iterative deepening a* search'],\n",
       " 'high throughput imaging': [],\n",
       " 'hill climbing search': [],\n",
       " 'hinge loss': [],\n",
       " 'housing': [],\n",
       " 'huber loss': [],\n",
       " 'human computer interaction': [],\n",
       " 'identity function': [],\n",
       " 'image processing': ['digital image processing'],\n",
       " 'income': ['tax'],\n",
       " 'infrastructure': [],\n",
       " 'intelligent water drops algorithm': [],\n",
       " 'intrusion detection': [],\n",
       " 'iterative deepening depth first search': [],\n",
       " 'journalism': ['news'],\n",
       " 'kullback leibler divergence loss': [],\n",
       " 'l1 hinge error': [],\n",
       " 'l1 loss': [],\n",
       " 'l2 loss': [],\n",
       " 'law': [],\n",
       " 'levenberg marquardt backpropagation': [],\n",
       " 'life sciences': [],\n",
       " 'literature': [],\n",
       " 'local beam search': [],\n",
       " 'local search': ['hill climbing search',\n",
       "  'local beam search',\n",
       "  'simulated annealing',\n",
       "  'travelling salesman problem'],\n",
       " 'logic based reasoning': ['first order logic',\n",
       "  'logical reasoning',\n",
       "  'modal logic',\n",
       "  'non monotonic logic',\n",
       "  'proportional logic'],\n",
       " 'logistic function': ['sigmoid function', 'softstep function'],\n",
       " 'manufacturing': [],\n",
       " 'margin classifier': [],\n",
       " 'marketing': [],\n",
       " 'mathematics': ['probability', 'statistics'],\n",
       " 'maxout function': [],\n",
       " 'mean bias error': [],\n",
       " 'media': ['digital media', 'journalism', 'multimedia'],\n",
       " 'medicine': ['epidemiology'],\n",
       " 'meteorology': ['climate', 'geographic information system (gis)'],\n",
       " 'microbiology': [],\n",
       " 'migrating birds optimization': [],\n",
       " 'mini max algorithm': [],\n",
       " 'mini batch gradient descent': [],\n",
       " 'miscellaneous ai applications': ['agriculture',\n",
       "  'assistance',\n",
       "  'astronomy',\n",
       "  'big data',\n",
       "  'crime',\n",
       "  'culture and humanities',\n",
       "  'data analytics',\n",
       "  'digital preservation',\n",
       "  'disaster management',\n",
       "  'law',\n",
       "  'literature',\n",
       "  'media',\n",
       "  'meteorology',\n",
       "  'nature and life sciences',\n",
       "  'politics',\n",
       "  'security',\n",
       "  'sociology',\n",
       "  'survey',\n",
       "  'urban development',\n",
       "  'imaging',\n",
       "  'satellite imaging/satellite imagery'],\n",
       " 'molecular docking': [],\n",
       " 'molecular biology': ['molecular docking'],\n",
       " 'movies': ['film'],\n",
       " 'multi class classification loss functions': ['kullback leibler divergence loss',\n",
       "  'multi class cross entropy loss'],\n",
       " 'multi class cross entropy loss': [],\n",
       " 'multiclass classification': [],\n",
       " 'music': [],\n",
       " 'nature and life sciences': ['biology',\n",
       "  'life sciences',\n",
       "  'environmental',\n",
       "  'microscopy'],\n",
       " 'negative log likelihood': ['cross entropy loss'],\n",
       " 'nesterov accelerated gradient': [],\n",
       " 'news': [],\n",
       " 'nouvelle ai': [],\n",
       " 'optical imagery': [],\n",
       " 'optimization algorithms': ['ant colony optimization',\n",
       "  'artificial bee colony optimization',\n",
       "  'artificial fish swarm algorithm',\n",
       "  'bacterial foraging optimization',\n",
       "  'first order optimization algorithms',\n",
       "  'flower pollination optimization',\n",
       "  'genetic algorithm optimization',\n",
       "  'intelligent water drops algorithm',\n",
       "  'migrating birds optimization',\n",
       "  'mini max algorithm',\n",
       "  'particle swarm optimization',\n",
       "  'second order optimization algorithms'],\n",
       " 'particle swarm optimization': [],\n",
       " 'performance metrics': ['accuracy metrics'],\n",
       " 'personal assistant': [],\n",
       " 'pharmaceutical': [],\n",
       " 'photography': [],\n",
       " 'politics': [],\n",
       " 'probability': [],\n",
       " 'pure heuristic search': [],\n",
       " 'quadratic loss': [],\n",
       " 'ramp function': [],\n",
       " 'recommendation systems': ['collaborative filtering',\n",
       "  'content based filtering',\n",
       "  'session based recommendations'],\n",
       " 'regression loss functions': ['absolute error loss',\n",
       "  'huber loss',\n",
       "  'mean bias error',\n",
       "  'squared error loss'],\n",
       " 'remote sensing': [],\n",
       " 'revenue forecast': [],\n",
       " 'sq rbf function': [],\n",
       " 'sqnl function': [],\n",
       " 'search algorithms': ['blind search',\n",
       "  'gravitational search algorithm',\n",
       "  'heuristic search',\n",
       "  'local search',\n",
       "  'tabu search algorithm'],\n",
       " 'second order optimization algorithms': [],\n",
       " 'security': ['cyber security', 'intrusion detection', 'surveillance'],\n",
       " 'sigmoid function': ['sigmoidal function'],\n",
       " 'sigmoidal function': ['binary sigmoidal function',\n",
       "  'bipolar sigmoidal function'],\n",
       " 'simulated annealing': [],\n",
       " 'sinc function': [],\n",
       " 'sinusoid function': [],\n",
       " 'sociology': [],\n",
       " 'softexponential function': [],\n",
       " 'softplus function': [],\n",
       " 'soft clipping function': [],\n",
       " 'soft margin classifier': [],\n",
       " 'softmax function': [],\n",
       " 'softsign function': [],\n",
       " 'softstep function': [],\n",
       " 'speech denoising': [],\n",
       " 'speech processing': [],\n",
       " 'sports': [],\n",
       " 'squared error loss': ['mean square error (mse)'],\n",
       " 'statistical ai': [],\n",
       " 'statistics': [],\n",
       " 'stochastic gradient descent': [],\n",
       " 'subsumption architecture': [],\n",
       " 'surveillance': [],\n",
       " 'survey': ['census'],\n",
       " 'tabu search algorithm': [],\n",
       " 'tanh function': [],\n",
       " 'tax': [],\n",
       " 'terrain mapping': [],\n",
       " 'text mining': [],\n",
       " 'trade': [],\n",
       " 'traffic': [],\n",
       " 'transfer learning': [],\n",
       " 'transport': [],\n",
       " 'transportation': ['automobiles',\n",
       "  'autonomous vehicles',\n",
       "  'traffic',\n",
       "  'transport'],\n",
       " 'travelling salesman problem': [],\n",
       " 'uniform cost search': [],\n",
       " 'urban development': ['architecture',\n",
       "  'banking',\n",
       "  'construction',\n",
       "  'demographics',\n",
       "  'design and modelling',\n",
       "  'education',\n",
       "  'housing',\n",
       "  'income',\n",
       "  'infrastructure',\n",
       "  'transportation'],\n",
       " 'variance based thresholding': [],\n",
       " 'verbal assistant': [],\n",
       " 'video games': [],\n",
       " 'visual arts': [],\n",
       " 'weather': [],\n",
       " 'aerial imagery': [],\n",
       " 'culture': [],\n",
       " 'disaster response': [],\n",
       " 'emissions': [],\n",
       " 'environmental': ['geography',\n",
       "  'emissions',\n",
       "  'pollution',\n",
       "  'quality',\n",
       "  'radiations'],\n",
       " 'forecast system': [],\n",
       " 'genes prediction': [],\n",
       " 'imaging': [],\n",
       " 'microscopy': [],\n",
       " 'multimedia': [],\n",
       " 'neuro biology': [],\n",
       " 'neuro imaging': [],\n",
       " 'neuroscience': ['neuro biology', 'neuro imaging'],\n",
       " 'pollution': [],\n",
       " 'quality': [],\n",
       " 'radiations': [],\n",
       " 'warning systems': [],\n",
       " 'a* search': [],\n",
       " 'adaptive piecewise linear (apl)': [],\n",
       " 'binary cross entropy (bce)': [],\n",
       " 'bipolar rectified linear unit (brelu)': [],\n",
       " 'exponential linear unit (elu)': [],\n",
       " 'gaussian error linear units (gelu)': [],\n",
       " 'geographic information system (gis)': [],\n",
       " 'geoinformatics (gis)': [],\n",
       " 'inverse square root linear unit (isrlu)': [],\n",
       " 'inverse square root unit (isru)': [],\n",
       " 'iterative deepening a* search': [],\n",
       " 'leaky rectified linear unit (leaky relu)': [],\n",
       " 'mean absolute error (mae)': ['l1 loss'],\n",
       " 'mean square error (mse)': ['l2 loss', 'quadratic loss'],\n",
       " 'parameteric rectified linear unit (prelu)': [],\n",
       " 'randomized leaky rectified linear unit (rrelu)': [],\n",
       " 'rectified linear unit (relu)': [],\n",
       " 'resilient backpropagation (rprop)': [],\n",
       " 's shaped rectified linear activation unit (srelu)': [],\n",
       " 'satellite imaging/satellite imagery': ['geospatial',\n",
       "  'optical imagery',\n",
       "  'remote sensing',\n",
       "  'terrain mapping',\n",
       "  'aerial imagery',\n",
       "  'geoinformatics (gis)'],\n",
       " 'scaled exponential linear unit (selu)': [],\n",
       " 'sigmoid linear unit (silu)': [],\n",
       " 'ahe': [],\n",
       " 'aiml': [],\n",
       " 'ai approaches': ['cybernetics and brain simulation',\n",
       "  'statistical learning',\n",
       "  'sub symbolic based ai',\n",
       "  'symbolic based ai'],\n",
       " 'ai concepts': ['ai software library',\n",
       "  'affective computing',\n",
       "  'evolutionary algorithms',\n",
       "  'game theory',\n",
       "  'human computer interaction',\n",
       "  'knowledge engineering',\n",
       "  'ontology',\n",
       "  'sequential learning',\n",
       "  'swarm intelligence',\n",
       "  'transfer learning'],\n",
       " 'ai methodoogy': ['activation functions',\n",
       "  'backpropogation',\n",
       "  'performance metrics',\n",
       "  'active learning',\n",
       "  'automl',\n",
       "  'bayesian inference',\n",
       "  'combinatorial optimization',\n",
       "  'data augmentation',\n",
       "  'density estimation',\n",
       "  'distributional reinforcement learning',\n",
       "  'domain adaptation',\n",
       "  'entity embeddings',\n",
       "  'feature importance',\n",
       "  'feature selection',\n",
       "  'hierarchical reinforcement learning',\n",
       "  'imitation learning',\n",
       "  'interpretable machine learning',\n",
       "  'loss function',\n",
       "  'meta learning',\n",
       "  'metric learning',\n",
       "  'model compression',\n",
       "  'model selection',\n",
       "  'multi goal reinforcement learning',\n",
       "  'multi agent reinforcement learning',\n",
       "  'multiple instance learning',\n",
       "  'network pruning',\n",
       "  'outlier detection',\n",
       "  'policy gradient methods',\n",
       "  'quantization',\n",
       "  'recommendation systems',\n",
       "  'representation learning',\n",
       "  'sparse learning',\n",
       "  'stochastic optimization',\n",
       "  'transfer learning'],\n",
       " 'ai ontology': ['artificial intelligence'],\n",
       " 'ai programming language': ['aiml',\n",
       "  'cplusplus',\n",
       "  'c language',\n",
       "  'centura',\n",
       "  'haskell',\n",
       "  'java',\n",
       "  'julia',\n",
       "  'lisp',\n",
       "  'matlab',\n",
       "  'perl',\n",
       "  'prolog',\n",
       "  'python',\n",
       "  'r language',\n",
       "  'ruby'],\n",
       " 'ai software library': ['ai programming language',\n",
       "  'azure',\n",
       "  'accord.net',\n",
       "  'apachemahout',\n",
       "  'caffe',\n",
       "  'dllearner',\n",
       "  'elki',\n",
       "  'eviews',\n",
       "  'encog',\n",
       "  'fama',\n",
       "  'h2o',\n",
       "  'ibmminer',\n",
       "  'knime',\n",
       "  'kxen',\n",
       "  'keras',\n",
       "  'lionsolver',\n",
       "  'liblinear',\n",
       "  'libsvm',\n",
       "  'mlpack',\n",
       "  'moa',\n",
       "  'mx net',\n",
       "  'mathematica',\n",
       "  'microsoft cntk',\n",
       "  'montecarlomachinelearninglibrary',\n",
       "  'neurosolutions',\n",
       "  'octave',\n",
       "  'opencv',\n",
       "  'opennn',\n",
       "  'oracledatamining',\n",
       "  'orange',\n",
       "  'pytorch',\n",
       "  'rcase',\n",
       "  'rapidminer',\n",
       "  'sap',\n",
       "  'sas',\n",
       "  'spss',\n",
       "  'sqlserveranalysisservices',\n",
       "  'statistica',\n",
       "  'shogun',\n",
       "  'spark mlib',\n",
       "  'stata',\n",
       "  'tensorflow',\n",
       "  'theano',\n",
       "  'weka',\n",
       "  'yale',\n",
       "  'yooreeka',\n",
       "  'mlpy',\n",
       "  'scikit learn'],\n",
       " 'atc': [],\n",
       " 'azure': [],\n",
       " 'accented speech recognition': [],\n",
       " 'accord.net': [],\n",
       " 'accuracy metrics': [],\n",
       " 'acoustic modelling': [],\n",
       " 'acoustic question answering': [],\n",
       " 'action classification': [],\n",
       " 'action detection': ['fine grained action detection',\n",
       "  'multiple action detection',\n",
       "  'skeleton based action recognition',\n",
       "  'temporal action proposal generation'],\n",
       " 'action localization': ['spatio temporal action localization'],\n",
       " 'active learning': ['known environment', 'unknown environment'],\n",
       " 'active object localization': [],\n",
       " 'activity recognition': ['egocentric activity recognition',\n",
       "  'group activity recognition',\n",
       "  'human activity recognition'],\n",
       " 'ad targeting': [],\n",
       " 'adaboost': [],\n",
       " 'adaptive thresholding': [],\n",
       " 'adversarial autoencoder': [],\n",
       " 'adversarial variational bayes': [],\n",
       " 'affective computing': [],\n",
       " 'affinity propogation': [],\n",
       " 'affix stemmers': [],\n",
       " 'agents and environment': [],\n",
       " 'agglomerative': [],\n",
       " 'alexnet': [],\n",
       " 'algorithmic trading': [],\n",
       " 'analogy': [],\n",
       " 'anamoly detection': ['elliptic envelope',\n",
       "  'local outlier factor',\n",
       "  'one class svm'],\n",
       " 'animation': [],\n",
       " 'anomaly detection': ['anomaly detection in surveillance videos',\n",
       "  'group anomaly detection',\n",
       "  'unsupervised anomaly detection',\n",
       "  'unsupervised anomaly detection in sound'],\n",
       " 'anomaly detection in surveillance videos': [],\n",
       " 'apachemahout': [],\n",
       " 'applications': ['arts and entertainment',\n",
       "  'health',\n",
       "  'miscellaneous ai applications',\n",
       "  'audit',\n",
       "  'automotive',\n",
       "  'finance and economics',\n",
       "  'marketing and advertising',\n",
       "  'military'],\n",
       " 'art analysis': [],\n",
       " 'articulatory synthesis': [],\n",
       " 'arts': ['linguistics', 'musical', 'visual and artistic', 'problem solving'],\n",
       " 'audio visual speech recognition': [],\n",
       " 'audio visual synchronization': [],\n",
       " 'audio visual video captioning': [],\n",
       " 'audit': [],\n",
       " 'automl': ['automated feature engineering',\n",
       "  'hyperparameter optimization',\n",
       "  'neural architecture search'],\n",
       " 'autoencoder': ['variational autoencoder'],\n",
       " 'autoencoders': [],\n",
       " 'automated feature engineering': [],\n",
       " 'automatic relevance determination': [],\n",
       " 'automatic summarization': [],\n",
       " 'automotive': ['manufacturing', 'transportation'],\n",
       " 'auxiliary classifier gan': [],\n",
       " 'auxiliary learning': [],\n",
       " 'backward feature elimination': [],\n",
       " 'basic image filtering': [],\n",
       " 'bayesian inference': ['probabilistic programming'],\n",
       " 'bayesian network': [],\n",
       " 'bayesian regression': ['automatic relevance determination',\n",
       "  'bayesian ridge regression'],\n",
       " 'bayesian ridge regression': [],\n",
       " 'belief network': [],\n",
       " 'bidirectional rnn': [],\n",
       " 'binary rbm with contrastive divergence': [],\n",
       " 'binary rbm with persistent contrastive divergence': [],\n",
       " 'binary shape analysis': ['boundary tracking',\n",
       "  'object labeling and counting',\n",
       "  'shape recognition',\n",
       "  'size filtering',\n",
       "  'skeletons and thinning'],\n",
       " 'birch': [],\n",
       " 'bird view synthesis': [],\n",
       " 'blind image deblurring': [],\n",
       " 'blind image quality assessment': [],\n",
       " 'boltzmann machine': ['dbm', 'rbm'],\n",
       " 'boosting': ['adaboost',\n",
       "  'brownboost',\n",
       "  'graddient boosting',\n",
       "  'lpboost',\n",
       "  'logitboost'],\n",
       " 'boundary detection': [],\n",
       " 'boundary equilibrium gan': [],\n",
       " 'boundary pattern analysis': [],\n",
       " 'boundary seeking gan': [],\n",
       " 'boundary tracking': [],\n",
       " 'brill tagger': [],\n",
       " 'brownboost': [],\n",
       " 'byte pair encoding': [],\n",
       " 'clahe': [],\n",
       " 'cnn': ['alexnet',\n",
       "  'lenet',\n",
       "  'mask rcnn',\n",
       "  'rcnn',\n",
       "  'resnet',\n",
       "  'vggnet',\n",
       "  'zfnet',\n",
       "  'googlenet/inception'],\n",
       " 'cplusplus': [],\n",
       " 'c language': [],\n",
       " 'caffe': [],\n",
       " 'camera localization': [],\n",
       " 'camera relocalization': [],\n",
       " 'car racing': [],\n",
       " 'catboost': [],\n",
       " 'centura': [],\n",
       " 'circle and ellipse detection': [],\n",
       " 'classification': ['multiclass classification',\n",
       "  'decision trees',\n",
       "  'gradient boost',\n",
       "  'logistic regression',\n",
       "  'naive bayes classifier',\n",
       "  'nearest neighbor',\n",
       "  'random forest',\n",
       "  'support vector machine'],\n",
       " 'classification methods': [],\n",
       " 'clinical training': [],\n",
       " 'clustering': ['affinity propogation',\n",
       "  'birch',\n",
       "  'density based spatial clustering',\n",
       "  'expectation maximization',\n",
       "  'hierarchical clustering',\n",
       "  'k means clustering',\n",
       "  'mean shift clustering',\n",
       "  'mixture model',\n",
       "  'optics algorithm',\n",
       "  'self organizing maps',\n",
       "  'spectral clustering'],\n",
       " 'clustering methods': [],\n",
       " 'co saliency detection': [],\n",
       " 'cognitive simulation': [],\n",
       " 'collaborative filtering': [],\n",
       " 'color constancy': ['few shot camera adaptive color constancy'],\n",
       " 'color image compression artifact reduction': [],\n",
       " 'colorization': [],\n",
       " 'combinatorial optimization': [],\n",
       " 'computational intelligence': [],\n",
       " 'computer aided interpretation of medical images': [],\n",
       " 'computer vision': ['image processing',\n",
       "  'intermediate level vision',\n",
       "  'low level vision',\n",
       "  'machine vision',\n",
       "  'neural network based',\n",
       "  'pattern recognition',\n",
       "  'tasks',\n",
       "  '3d vision'],\n",
       " 'concatenation synthesis': ['diphone synthesis',\n",
       "  'domain specific synthesis',\n",
       "  'unit selection synthesis'],\n",
       " 'conditional image generation': [],\n",
       " 'conditional planning': [],\n",
       " 'conditional vae': [],\n",
       " 'conditional random fields approach': [],\n",
       " 'constrained diffeomorphic image registration': [],\n",
       " 'content based filtering': [],\n",
       " 'content based image retrieval': [],\n",
       " 'content generation': [],\n",
       " 'context based emotion recognition': [],\n",
       " 'continuous control': [],\n",
       " 'continuous object recognition': [],\n",
       " 'contour detection': [],\n",
       " 'conversational speech recognition': [],\n",
       " 'converssational emotion recognition': [],\n",
       " 'convolutional neural networks': [],\n",
       " 'coreference resolution': [],\n",
       " 'corner and interest point detection': ['local invariant feature detectors and descriptors'],\n",
       " 'coupled gan': [],\n",
       " 'curved text detection': [],\n",
       " 'cybernetics and brain simulation': [],\n",
       " 'dbm': [],\n",
       " 'dllearner': [],\n",
       " 'data augmentation': [],\n",
       " 'deblurring': ['blind image deblurring', 'single image blind deblurring'],\n",
       " 'decision tree regressor': [],\n",
       " 'decision stump': [],\n",
       " 'decision theory': [],\n",
       " 'decision trees': ['decision stump'],\n",
       " 'deep adversarial networks': [],\n",
       " 'deep attention': [],\n",
       " 'deep learning': ['transfer learning',\n",
       "  'convolutional neural networks',\n",
       "  'recurrent neural networks',\n",
       "  'recursive neural networks',\n",
       "  'unsupervised pretrained networks (upn)'],\n",
       " 'deep rnn': [],\n",
       " 'deep reinforcement learning': [],\n",
       " 'deep feedforward and recurrent neural networks': [],\n",
       " 'deep learning synthesis': [],\n",
       " 'default reasoning and the qualification problem': [],\n",
       " 'defocus estimation': [],\n",
       " 'deformable object manipulation': [],\n",
       " 'demosaicking': [],\n",
       " 'denoising': ['speech denoising', 'image denoising'],\n",
       " 'denoising vae': [],\n",
       " 'dense pixel correspondence estimation': [],\n",
       " 'dense video captioning': [],\n",
       " 'density based spatial clustering': [],\n",
       " 'density estimation': [],\n",
       " 'depiction invariant object recognition': [],\n",
       " 'depth and camera motion': [],\n",
       " 'depth completion': [],\n",
       " 'depth estimation': ['depth and camera motion',\n",
       "  'depth prediction for outdoor scene',\n",
       "  'monocular depth estimation',\n",
       "  'stereo depth estimation'],\n",
       " 'depth map super resolution': [],\n",
       " 'depth prediction for outdoor scene': [],\n",
       " 'description logic': [],\n",
       " 'design treatment plans': [],\n",
       " 'dictionary based mt': [],\n",
       " 'digit recognition': [],\n",
       " 'digital image processing': ['ahe', 'clahe', 'gcn'],\n",
       " 'dimensionality reduction': ['autoencoder',\n",
       "  'backward feature elimination',\n",
       "  'factor analysis',\n",
       "  'forward feature selection',\n",
       "  'high correlation filter',\n",
       "  'ica',\n",
       "  'low variance filter',\n",
       "  'methods based on projections',\n",
       "  'missing value ratio',\n",
       "  'pca',\n",
       "  'projection pursuit',\n",
       "  'svd',\n",
       "  'sammon mapping',\n",
       "  'umap',\n",
       "  't sne',\n",
       "  'flexible discriminant analysis (fda)',\n",
       "  'mixture discriminant analysis (mda)',\n",
       "  'multidimensional scaling (mds)',\n",
       "  'partial least squares regression (plsr)',\n",
       "  'principal component regression (pcr)',\n",
       "  'quadratic discriminant analysis (qda)'],\n",
       " 'dimensionality reduction methods': [],\n",
       " 'diphone synthesis': [],\n",
       " 'discogan': [],\n",
       " 'discourse': ['automatic summarization', 'coreference resolution'],\n",
       " 'disease prediction and diagnosis': ['cancer', 'diabetes'],\n",
       " 'disparity estimation': [],\n",
       " 'distant speech recognition': [],\n",
       " 'distributional reinforcement learning': [],\n",
       " 'distributional semantics': ['hyperspace analogue to language',\n",
       "  'latent semantic analysis',\n",
       "  'random indexing',\n",
       "  'semantic folding'],\n",
       " 'distributional learning': [],\n",
       " 'divisive': [],\n",
       " 'document embedding': [],\n",
       " 'document image classification': [],\n",
       " 'document to image conversion': [],\n",
       " 'domain specific synthesis': [],\n",
       " 'domain adaptation': ['domain generalization',\n",
       "  'partial domain adaptation',\n",
       "  'predictive domain adaptation',\n",
       "  'unsupervised domain adaptation'],\n",
       " 'domain generalization': [],\n",
       " 'drawing pictures': [],\n",
       " 'drug creation': ['molecular docking', 'pharmaceutical'],\n",
       " 'dualgan': [],\n",
       " 'dynamic belief networks': [],\n",
       " 'dynamic decision networks': [],\n",
       " 'dynamic pricing': [],\n",
       " 'dynamic texture recognition': [],\n",
       " 'dynamic time warping': [],\n",
       " 'elki': [],\n",
       " 'eviews': [],\n",
       " 'edge detection': [],\n",
       " 'effector': [],\n",
       " 'egocentric activity recognition': [],\n",
       " 'elastic net': [],\n",
       " 'elliptic envelope': [],\n",
       " 'embodied question answering': [],\n",
       " 'embodied intelligence': [],\n",
       " 'emotion recognition': ['context based emotion recognition',\n",
       "  'converssational emotion recognition',\n",
       "  'multimodal emotion recognition',\n",
       "  'speech emotion recognition'],\n",
       " 'encog': [],\n",
       " 'energy based gan': [],\n",
       " 'ensemble learning': ['boosting',\n",
       "  'random forest',\n",
       "  'stacking',\n",
       "  'bootstrap aggregating (bagging)'],\n",
       " 'entity embeddings': [],\n",
       " 'event based vision': [],\n",
       " 'evolutionary algorithms': ['genetic algorithms'],\n",
       " 'example based mt': [],\n",
       " 'expectation maximization': [],\n",
       " 'expressive operator descriptions': [],\n",
       " 'fama': [],\n",
       " 'face alignment': [],\n",
       " 'face detection': [],\n",
       " 'face generation': [],\n",
       " 'face recognition': [],\n",
       " 'face reconstruction': [],\n",
       " 'face transfer': [],\n",
       " 'face verification': [],\n",
       " 'facial expression recognition': [],\n",
       " 'facial inpainting': [],\n",
       " 'facial recognition and modelling': ['face alignment',\n",
       "  'face detection',\n",
       "  'face generation',\n",
       "  'face recognition',\n",
       "  'face reconstruction',\n",
       "  'face verification',\n",
       "  'facial expression recognition'],\n",
       " 'factor analysis': [],\n",
       " 'factual visual question answering': [],\n",
       " 'fake image detection': [],\n",
       " 'feature importance': [],\n",
       " 'feature learning': ['deep learning', 'deep reinforcement learning'],\n",
       " 'feature selection': [],\n",
       " 'few shot camera adaptive color constancy': [],\n",
       " 'few shot image classification': [],\n",
       " 'few shot imitation learning': [],\n",
       " 'few shot learning': ['few shot camera adaptive color constancy',\n",
       "  'few shot imitation learning',\n",
       "  'few shot relation classification'],\n",
       " 'few shot relation classification': [],\n",
       " 'finance and economics': ['business',\n",
       "  'economics',\n",
       "  'finance',\n",
       "  'algorithmic trading',\n",
       "  'fraud detection',\n",
       "  'market analysis and data mining',\n",
       "  'personal finance',\n",
       "  'portfolio management',\n",
       "  'underwriting'],\n",
       " 'fine grained action detection': [],\n",
       " 'fine grained image classification': [],\n",
       " 'fine grained image recognition': [],\n",
       " 'fine grain image categorization': [],\n",
       " 'first order logic': [],\n",
       " 'forgery': [],\n",
       " 'formant synthesis': [],\n",
       " 'forward feature selection': [],\n",
       " 'frame systems and semantic networks': [],\n",
       " 'fraud detection': [],\n",
       " 'fraudalent ads': [],\n",
       " 'fundamental': ['distributed artificial intelligence',\n",
       "  'mathematics',\n",
       "  'ai approaches',\n",
       "  'ai concepts',\n",
       "  'ai methodoogy',\n",
       "  'general intelligence',\n",
       "  'knowledge and reasoning',\n",
       "  'machine learning',\n",
       "  'planning',\n",
       "  'problem solving',\n",
       "  'search and optimization',\n",
       "  'social intelligence',\n",
       "  'uncertain knowledge and reasoning'],\n",
       " 'gan': ['auxiliary classifier gan',\n",
       "  'boundary equilibrium gan',\n",
       "  'boundary seeking gan',\n",
       "  'coupled gan',\n",
       "  'discogan',\n",
       "  'dualgan',\n",
       "  'energy based gan',\n",
       "  'generative adversarial parallelization',\n",
       "  'gibbsnet',\n",
       "  'improved training for wasserstein gan',\n",
       "  'infogan',\n",
       "  'least squares gan',\n",
       "  'magan',\n",
       "  'mode regularized gan',\n",
       "  'softmax gan',\n",
       "  'vanilla gan',\n",
       "  'wasserstein gan',\n",
       "  'f gan',\n",
       "  'adversarial feature learning & adversarially learned inference',\n",
       "  'conditional gan (cgan)',\n",
       "  'deep convolutional gan (dcgan)',\n",
       "  'laplacian pyramid gan (lapgan)',\n",
       "  'super resolution gan (srgan)'],\n",
       " 'gbm': [],\n",
       " 'gcn': [],\n",
       " 'game theory': [],\n",
       " 'games': ['sports', 'video games'],\n",
       " 'gated recurrent unit': ['gru'],\n",
       " 'gaze estimation': [],\n",
       " 'general intelligence': ['agents and environment'],\n",
       " 'general logical descriptions': [],\n",
       " 'general ontology': [],\n",
       " 'generalized zero shot learning': [],\n",
       " 'generative adversarial networks': ['gan'],\n",
       " 'generative adversarial parallelization': [],\n",
       " 'genetic algorithms': [],\n",
       " 'gesture recognition': ['hand gesture recognition',\n",
       "  'isolated gestures recognition'],\n",
       " 'gibbsnet': [],\n",
       " 'graddient boosting': [],\n",
       " 'gradient boost regressor': [],\n",
       " 'gradient boost': ['catboost',\n",
       "  'catboost',\n",
       "  'gbm',\n",
       "  'lightgbm',\n",
       "  'light gbm',\n",
       "  'xgboost',\n",
       "  'xgboost'],\n",
       " 'grammar induction': ['distributional learning',\n",
       "  'grammatical inference by genetic algorithms',\n",
       "  'grammatical inference by greedy algorithms',\n",
       "  'grammatical inference by trial and error',\n",
       "  'pattern theory'],\n",
       " 'grammatical inference by genetic algorithms': [],\n",
       " 'grammatical inference by greedy algorithms': ['byte pair encoding',\n",
       "  'lempel ziv welch algorithm',\n",
       "  'sequitur'],\n",
       " 'grammatical inference by trial and error': [],\n",
       " 'graph embedding': [],\n",
       " 'graph representation learning': [],\n",
       " 'group activity recognition': [],\n",
       " 'group anomaly detection': [],\n",
       " 'h2o': [],\n",
       " 'hmm based synthesis': [],\n",
       " 'hand gesture recognition': [],\n",
       " 'handwritten digit recognition': [],\n",
       " 'haskell': [],\n",
       " 'health care': ['medicine',\n",
       "  'clinical training',\n",
       "  'computer aided interpretation of medical images',\n",
       "  'design treatment plans',\n",
       "  'disease prediction and diagnosis',\n",
       "  'drug creation',\n",
       "  'medication management',\n",
       "  'provide consultations and companion robots'],\n",
       " 'hebbian learning': [],\n",
       " 'hidden markov models based': [],\n",
       " 'hidden markov models': [],\n",
       " 'hierarchical clustering': ['agglomerative', 'divisive'],\n",
       " 'hierarchical decomposition': [],\n",
       " 'hierarchical reinforcement learning': [],\n",
       " 'high correlation filter': [],\n",
       " 'homography estimation': [],\n",
       " 'hopfield network': [],\n",
       " 'hough transform': [],\n",
       " 'huber regression': [],\n",
       " 'human object interaction detection': ['human computer interaction'],\n",
       " 'human activity recognition': [],\n",
       " 'hybrid mt': [],\n",
       " 'hyperparameter optimization': [],\n",
       " 'hyperspace analogue to language': [],\n",
       " 'hyperspectral': ['hyperspectral image segmentation',\n",
       "  'hyperspectral unmixing'],\n",
       " 'hyperspectral image classification': [],\n",
       " 'hyperspectral image segmentation': [],\n",
       " 'hyperspectral unmixing': [],\n",
       " 'hypothesis of creative patterns': [],\n",
       " 'ibmminer': [],\n",
       " 'ica': [],\n",
       " 'image based localization': [],\n",
       " 'image to image translation': ['bird view synthesis',\n",
       "  'keypoint localization',\n",
       "  'optical character recognition',\n",
       "  'unsupervised image to image translation'],\n",
       " 'image captioning': [],\n",
       " 'image classification': ['few shot image classification',\n",
       "  'fine grained image classification',\n",
       "  'hyperspectral image classification',\n",
       "  'semi supervised image classification'],\n",
       " 'image clustering': [],\n",
       " 'image comprehension': [],\n",
       " 'image compression': ['color image compression artifact reduction',\n",
       "  'jpeg compression artifact reduction',\n",
       "  'lossy compression artifact reduction'],\n",
       " 'image cropping': [],\n",
       " 'image declipping': [],\n",
       " 'image deconvolution': [],\n",
       " 'image dehazing': [],\n",
       " 'image denoising': ['salt and pepper noise removal'],\n",
       " 'image enhancement': ['local color enhancement',\n",
       "  'low light image enhancement'],\n",
       " 'image generation': [],\n",
       " 'image imputation': [],\n",
       " 'image inpainting': ['facial inpainting'],\n",
       " 'image manipulation detection': [],\n",
       " 'image matching': [],\n",
       " 'image matting': [],\n",
       " 'image quality assessment': ['blind image quality assessment',\n",
       "  'no reference image quality assessment'],\n",
       " 'image recognition': ['fine grained image recognition'],\n",
       " 'image reconstruction': [],\n",
       " 'image rectification': [],\n",
       " 'image registration': ['constrained diffeomorphic image registration'],\n",
       " 'image restoration': [],\n",
       " 'image retrieval': ['content based image retrieval',\n",
       "  'multi label image retrieval',\n",
       "  'sketch based image retrieval'],\n",
       " 'image similarity search': [],\n",
       " 'image steganography': [],\n",
       " 'image stitching': [],\n",
       " 'image stylization': [],\n",
       " 'image super resolution': [],\n",
       " 'image transformations and camera calibration': ['image rectification',\n",
       "  '3 d reconstruction'],\n",
       " 'image categorization': ['fine grain image categorization'],\n",
       " 'imitation learning': [],\n",
       " 'improved training for wasserstein gan': [],\n",
       " 'inductive learning': [],\n",
       " 'infogan': [],\n",
       " 'information theory': [],\n",
       " 'infrared and visible image fusion': [],\n",
       " 'instance search': [],\n",
       " 'instance segmentation': ['3d instance segmentation',\n",
       "  '3d semantic instance segmentation'],\n",
       " 'intelligent agent': [],\n",
       " 'intelligent surveillance': [],\n",
       " 'intent segmentation': [],\n",
       " 'interactive segmentation': [],\n",
       " 'interlingual mt': [],\n",
       " 'intermediate level vision': ['binary shape analysis',\n",
       "  'boundary pattern analysis',\n",
       "  'circle and ellipse detection',\n",
       "  'hough transform',\n",
       "  'line detection',\n",
       "  'pattern matching'],\n",
       " 'interpretable machine learning': [],\n",
       " 'intrinsic image decomposition': [],\n",
       " 'invariants and perspective': [],\n",
       " 'isolated gestures recognition': [],\n",
       " 'java': [],\n",
       " 'joke generation': [],\n",
       " 'jpeg compression artifact reduction': [],\n",
       " 'julia': [],\n",
       " 'k means clustering': [],\n",
       " 'knime': [],\n",
       " 'kxen': [],\n",
       " 'keras': [],\n",
       " 'kernel methods': ['svm'],\n",
       " 'kernel svm': [],\n",
       " 'keypoint localization': [],\n",
       " 'keyword spotting': [],\n",
       " 'knowledge engineering': [],\n",
       " 'knowledge graph embedding': [],\n",
       " 'knowledge and reasoning': ['logic based reasoning',\n",
       "  'default reasoning and the qualification problem',\n",
       "  'general ontology'],\n",
       " 'knowledge based': ['reinforcement learning'],\n",
       " 'known environment': [],\n",
       " 'lars lasso': [],\n",
       " 'lionsolver': [],\n",
       " 'lpboost': [],\n",
       " 'lstm': [],\n",
       " 'language based temporal localization': [],\n",
       " 'large scale person re identification': [],\n",
       " 'large vocabulary continuous speech recognition': [],\n",
       " 'lasso regression': [],\n",
       " 'latent semantic analysis': [],\n",
       " 'lenet': [],\n",
       " 'learning from observations': ['decision trees',\n",
       "  'general logical descriptions',\n",
       "  'inductive learning',\n",
       "  'information theory'],\n",
       " 'least angle regression': [],\n",
       " 'least squares gan': [],\n",
       " 'lemmatisation algorithms': [],\n",
       " 'lemmatization': [],\n",
       " 'lempel ziv welch algorithm': [],\n",
       " 'lexeme based morphology': [],\n",
       " 'lexical semantics': [],\n",
       " 'lexical recognition': [],\n",
       " 'liblinear': [],\n",
       " 'libsvm': [],\n",
       " 'lightgbm': [],\n",
       " 'light gbm': [],\n",
       " 'line detection': [],\n",
       " 'line segment detection': [],\n",
       " 'linear regression': ['elastic net',\n",
       "  'elastic net',\n",
       "  'ordinary least squares regression',\n",
       "  'stochastic gradient descent regression'],\n",
       " 'linguistics': ['analogy',\n",
       "  'hypothesis of creative patterns',\n",
       "  'joke generation',\n",
       "  'metaphor and simile',\n",
       "  'neologism',\n",
       "  'poetry',\n",
       "  'story generation'],\n",
       " 'lipreading': [],\n",
       " 'lisp': [],\n",
       " 'local color enhancement': [],\n",
       " 'local invariant feature detectors and descriptors': [],\n",
       " 'local outlier factor': [],\n",
       " 'logic based': [],\n",
       " 'logical reasoning': ['description logic',\n",
       "  'frame systems and semantic networks',\n",
       "  'indexing, retrieval and unification'],\n",
       " 'logistic regression': [],\n",
       " 'logitboost': [],\n",
       " 'loss function': ['classification loss functions',\n",
       "  'embedding loss functions',\n",
       "  'regression loss functions'],\n",
       " 'lossy compression artifact reduction': [],\n",
       " 'low level vision': ['basic image filtering',\n",
       "  'corner and interest point detection',\n",
       "  'edge detection',\n",
       "  'mathematical morphology',\n",
       "  'texture',\n",
       "  'thresholding'],\n",
       " 'low light image enhancement': [],\n",
       " 'low variance filter': [],\n",
       " 'magan': [],\n",
       " 'matlab': [],\n",
       " 'mlpack': [],\n",
       " 'moa': [],\n",
       " 'mx net': [],\n",
       " 'machine learning': ['data science',\n",
       "  'feature engineering',\n",
       "  'belief network',\n",
       "  'feature learning',\n",
       "  'knowledge based',\n",
       "  'learning from observations',\n",
       "  'neural networks',\n",
       "  'semi supervised learning',\n",
       "  'supervised learning',\n",
       "  'unsupervised learning'],\n",
       " 'machine translation': ['example based mt',\n",
       "  'hybrid mt',\n",
       "  'neural mt',\n",
       "  'rule based mt',\n",
       "  'statistical based mt'],\n",
       " 'machine vision': [],\n",
       " 'manner of articulation detection': [],\n",
       " 'markerless motion capture': [],\n",
       " 'market analysis and data mining': [],\n",
       " 'marketing forecasting': [],\n",
       " 'marketing and advertising': ['marketing',\n",
       "  'trade',\n",
       "  'ad targeting',\n",
       "  'content generation',\n",
       "  'dynamic pricing',\n",
       "  'fraudalent ads',\n",
       "  'marketing forecasting',\n",
       "  'predictive analytics for marketing',\n",
       "  'predictive customer service',\n",
       "  'products recognition',\n",
       "  'programmatic advertising',\n",
       "  'recommendation engine'],\n",
       " 'mask rcnn': [],\n",
       " 'matching algorithms': [],\n",
       " 'material classification': ['materials screening'],\n",
       " 'material recognition': [],\n",
       " 'materials screening': [],\n",
       " 'mathematica': [],\n",
       " 'mathematical morphology': [],\n",
       " 'mathematical question answering': [],\n",
       " 'mean shift clustering': [],\n",
       " 'medication management': [],\n",
       " 'meta learning': [],\n",
       " 'metamerism': [],\n",
       " 'metaphor and simile': [],\n",
       " 'methods based on projections': [],\n",
       " 'metric learning': [],\n",
       " 'microsoft cntk': [],\n",
       " 'military': [],\n",
       " 'missing markers reconstruction': [],\n",
       " 'missing value ratio': [],\n",
       " 'mixture model': [],\n",
       " 'modal logic': [],\n",
       " 'mode regularized gan': [],\n",
       " 'model compression': [],\n",
       " 'model selection': [],\n",
       " 'monocular depth estimation': [],\n",
       " 'montecarlomachinelearninglibrary': [],\n",
       " 'morpheme based morphology': [],\n",
       " 'morphological segmentation': ['lexeme based morphology',\n",
       "  'morpheme based morphology',\n",
       "  'word based morphology'],\n",
       " 'motion': ['optical flow',\n",
       "  'stereo from motion',\n",
       "  'the kalman filter',\n",
       "  'time to adjacency analysis',\n",
       "  'wide baseline matching'],\n",
       " 'motion capture': ['markerless motion capture',\n",
       "  'missing markers reconstruction'],\n",
       " 'motion estimation': [],\n",
       " 'motion planning': [],\n",
       " 'motion segmentation': [],\n",
       " 'multi goal reinforcement learning': [],\n",
       " 'multi label image retrieval': [],\n",
       " 'multi layer perceptron': [],\n",
       " 'multi oriented scene text detection': [],\n",
       " 'multi task learning': [],\n",
       " 'multi agent reinforcement learning': [],\n",
       " 'multi modal image segmentation': [],\n",
       " 'multilayer feed forward networks': ['cnn',\n",
       "  'multi layer perceptron',\n",
       "  'radial basis network'],\n",
       " 'multimodal emotion recognition': [],\n",
       " 'multimodal unsupervised image to image translation': [],\n",
       " 'multiple action detection': [],\n",
       " 'multiple instance learning': [],\n",
       " 'multiple object tracking': [],\n",
       " 'multiview learning': [],\n",
       " 'music genre transfer': [],\n",
       " 'musical': ['music'],\n",
       " 'naive bayes classifier': [],\n",
       " 'named entity recognition': ['conditional random fields approach'],\n",
       " 'natural language generation': [],\n",
       " 'natural language processing': ['text mining',\n",
       "  'atc',\n",
       "  'discourse',\n",
       "  'semantics',\n",
       "  'speech',\n",
       "  'syntax'],\n",
       " 'natural language understanding': [],\n",
       " 'nearest neighbor': [],\n",
       " 'neologism': [],\n",
       " 'network embedding': [],\n",
       " 'network pruning': [],\n",
       " 'neural network based': ['deep learning'],\n",
       " 'neural architecture search': [],\n",
       " 'neural mt': [],\n",
       " 'neural networks': ['artificial neural networks'],\n",
       " 'neurosolutions': [],\n",
       " 'no reference image quality assessment': [],\n",
       " 'noisy speech recognition': [],\n",
       " 'non monotonic logic': [],\n",
       " 'non optimization based': ['k nn'],\n",
       " 'novel view synthesis': [],\n",
       " 'ocr': [],\n",
       " 'optics algorithm': [],\n",
       " 'object classification': [],\n",
       " 'object counting': [],\n",
       " 'object detection': ['salient object detection',\n",
       "  'weakly supervised object detection'],\n",
       " 'object labeling and counting': [],\n",
       " 'object localization': ['active object localization',\n",
       "  'image based localization',\n",
       "  'unsupervised object localization',\n",
       "  'weakly supervised object localization'],\n",
       " 'object recognition': ['continuous object recognition',\n",
       "  'depiction invariant object recognition',\n",
       "  '3d object recognition'],\n",
       " 'object reconstruction': [],\n",
       " 'object tracking': ['multiple object tracking',\n",
       "  'thermal infrared object tracking',\n",
       "  'visual object tracking'],\n",
       " 'observation completion': [],\n",
       " 'octave': [],\n",
       " 'one class svm': [],\n",
       " 'one shot learning': [],\n",
       " 'ontology': [],\n",
       " 'opencv': [],\n",
       " 'opennn': [],\n",
       " 'open domain question answering': [],\n",
       " 'optical character recognition': ['handwritten digit recognition'],\n",
       " 'optical flow': [],\n",
       " 'optical flow estimation': [],\n",
       " 'oracledatamining': [],\n",
       " 'orange': [],\n",
       " 'ordinary least squares regression': [],\n",
       " 'orthogonal matching pursuit': [],\n",
       " 'out of distribution detection': [],\n",
       " 'outlier detection': [],\n",
       " 'pca': [],\n",
       " 'parsing': [],\n",
       " 'partial order planning': [],\n",
       " 'partial domain adaptation': [],\n",
       " 'partially instantiated operators': [],\n",
       " 'passive aggressive algorithms': [],\n",
       " 'passive learning': ['known environment', 'unknown environment'],\n",
       " 'pattern matching': [],\n",
       " 'pattern recognition': ['kernel methods', 'statistical methods'],\n",
       " 'pattern theory': [],\n",
       " 'pedestrian detection': [],\n",
       " 'perception': ['computer vision', 'speech'],\n",
       " 'perceptrons': [],\n",
       " 'perl': [],\n",
       " 'person re identification': ['large scale person re identification',\n",
       "  'semi supervised person re identification',\n",
       "  'unsupervised person re identification',\n",
       "  'video based person re identification'],\n",
       " 'person retrieval': [],\n",
       " 'personal finance': [],\n",
       " 'phonotactic cues': [],\n",
       " 'photo to caricature translation': [],\n",
       " 'planning': ['conditional planning',\n",
       "  'expressive operator descriptions',\n",
       "  'hierarchical decomposition',\n",
       "  'intelligent agent',\n",
       "  'knowledge engineering',\n",
       "  'partial order planning',\n",
       "  'partially instantiated operators'],\n",
       " 'pos tagging': ['brill tagger', 'structure regularization method'],\n",
       " 'poetry': [],\n",
       " 'policy gradient methods': [],\n",
       " 'portfolio management': [],\n",
       " 'pose estimation': [],\n",
       " 'predictive analytics for marketing': [],\n",
       " 'predictive domain adaptation': [],\n",
       " 'predictive customer service': [],\n",
       " 'probabilistic programming': [],\n",
       " 'probabilistic methods for uncertain reasoning': [],\n",
       " 'problem solving': [],\n",
       " 'products recognition': [],\n",
       " 'programmatic advertising': [],\n",
       " 'projection pursuit': [],\n",
       " 'prolog': [],\n",
       " 'proportional logic': [],\n",
       " 'provide consultations and companion robots': [],\n",
       " 'python': [],\n",
       " 'pytorch': [],\n",
       " 'q learning': [],\n",
       " 'quantization': [],\n",
       " 'question answering': ['mathematical question answering',\n",
       "  'open domain question answering'],\n",
       " 'rbm': ['binary rbm with contrastive divergence',\n",
       "  'binary rbm with persistent contrastive divergence'],\n",
       " 'rcase': [],\n",
       " 'rcnn': [],\n",
       " 'rnn': ['bidirectional rnn', 'deep rnn', 'gated recurrent unit', 'lstm'],\n",
       " 'r language': [],\n",
       " 'radial basis network': [],\n",
       " 'radial basis svm': [],\n",
       " 'random forest': [],\n",
       " 'random indexing': [],\n",
       " 'rapidminer': [],\n",
       " 'recommendation engine': [],\n",
       " 'recurrent neural networks': [],\n",
       " 'recursive neural networks': [],\n",
       " 'regression': ['bayesian regression',\n",
       "  'decision tree regressor',\n",
       "  'gradient boost regressor',\n",
       "  'lars lasso',\n",
       "  'lasso regression',\n",
       "  'least angle regression',\n",
       "  'linear regression',\n",
       "  'orthogonal matching pursuit',\n",
       "  'passive aggressive algorithms',\n",
       "  'ridge regression',\n",
       "  'robustness regression'],\n",
       " 'reinforcement learning': ['active learning',\n",
       "  'deep adversarial networks',\n",
       "  'passive learning',\n",
       "  'q learning',\n",
       "  'temporal difference'],\n",
       " 'relationship extraction': [],\n",
       " 'representation learning': ['document embedding',\n",
       "  'graph embedding',\n",
       "  'graph representation learning',\n",
       "  'knowledge graph embedding',\n",
       "  'network embedding',\n",
       "  'sentence embeddings',\n",
       "  'unsupervised representation learning',\n",
       "  'word embeddings'],\n",
       " 'resnet': [],\n",
       " 'ridge regression': [],\n",
       " 'robot navigation': [],\n",
       " 'robot task planning': [],\n",
       " 'robotic grasping': [],\n",
       " 'robotics': ['deformable object manipulation',\n",
       "  'effector',\n",
       "  'motion planning',\n",
       "  'robot navigation',\n",
       "  'robot task planning',\n",
       "  'robotic grasping',\n",
       "  'sensor',\n",
       "  'visual navigation',\n",
       "  'visual odometry',\n",
       "  'language based assistance'],\n",
       " 'robust speech recognition': [],\n",
       " 'robustness regression': ['huber regression',\n",
       "  'theil sen estimator',\n",
       "  'random sample consensus(ransac)'],\n",
       " 'ruby': [],\n",
       " 'rule based mt': ['dictionary based mt',\n",
       "  'interlingual mt',\n",
       "  'transfer based mt'],\n",
       " 'sap': [],\n",
       " 'sas': [],\n",
       " 'sbd': [],\n",
       " 'spss': [],\n",
       " 'sqlserveranalysisservices': [],\n",
       " 'statistica': [],\n",
       " 'svd': [],\n",
       " 'svm': [],\n",
       " 'saliency detection': ['co saliency detection', 'video saliency detection'],\n",
       " 'saliency prediction': [],\n",
       " 'salient object detection': [],\n",
       " 'salt and pepper noise removal': [],\n",
       " 'sammon mapping': [],\n",
       " 'scene aware dialogue': [],\n",
       " 'scene classification': [],\n",
       " 'scene flow estimation': [],\n",
       " 'scene graph generation': [],\n",
       " 'scene parsing': ['scene graph generation',\n",
       "  'scene recognition',\n",
       "  'scene text recognition',\n",
       "  'scene understanding'],\n",
       " 'scene recognition': [],\n",
       " 'scene segmentation': [],\n",
       " 'scene text detection': ['curved text detection',\n",
       "  'multi oriented scene text detection'],\n",
       " 'scene text recognition': [],\n",
       " 'scene understanding': [],\n",
       " 'scruffy': [],\n",
       " 'search and optimization': ['optimization algorithms', 'search algorithms'],\n",
       " 'self organizing maps': [],\n",
       " 'semantic segmentation': ['weakly supervised semantic segmentation'],\n",
       " 'semantic folding': [],\n",
       " 'semantics': ['distributional semantics',\n",
       "  'lexical semantics',\n",
       "  'machine translation',\n",
       "  'named entity recognition',\n",
       "  'natural language generation',\n",
       "  'natural language understanding',\n",
       "  'ocr',\n",
       "  'question answering',\n",
       "  'relationship extraction',\n",
       "  'sentiment analysis',\n",
       "  'textual entailment recognition',\n",
       "  'topic segmentation',\n",
       "  'word sense disambiguation'],\n",
       " 'semi supervised image classification': [],\n",
       " 'semi supervised learning': [],\n",
       " 'semi supervised person re identification': [],\n",
       " 'semi supervised video object segmentation': [],\n",
       " 'sensor': [],\n",
       " 'sensor modeling': [],\n",
       " 'sentence embeddings': [],\n",
       " 'sentence segmentaion': [],\n",
       " 'sentiment analysis': [],\n",
       " 'sequence to sequence speecg recognition': [],\n",
       " 'sequential learning': [],\n",
       " 'sequitur': [],\n",
       " 'session based recommendations': [],\n",
       " 'shape recognition': [],\n",
       " 'shogun': [],\n",
       " 'sign language recognition': [],\n",
       " 'sign language translation': [],\n",
       " 'simultaneous localization and mapping': [],\n",
       " 'sinewave synthesis': [],\n",
       " 'single image blind deblurring': [],\n",
       " 'size filtering': [],\n",
       " 'skeleton based action recognition': [],\n",
       " 'skeletons and thinning': [],\n",
       " 'sketch based image retrieval': [],\n",
       " 'sketch recognition': [],\n",
       " 'social intelligence': [],\n",
       " 'soft computing': ['computational creativity',\n",
       "  'evolutionary computation',\n",
       "  'fuzzy systems'],\n",
       " 'softmax gan': [],\n",
       " 'spark mlib': [],\n",
       " 'sparse learning': [],\n",
       " 'spatio temporal action localization': [],\n",
       " 'speaker identification': [],\n",
       " 'speaker recognition': [],\n",
       " 'speaking style synthesis': [],\n",
       " 'spectral clustering': [],\n",
       " 'spectral estimation': ['spectrum cartography'],\n",
       " 'spectrum cartography': [],\n",
       " 'speech': ['speech processing',\n",
       "  'acoustic modelling',\n",
       "  'acoustic question answering',\n",
       "  'audio visual speech recognition',\n",
       "  'denoising',\n",
       "  'emotion recognition',\n",
       "  'keyword spotting',\n",
       "  'large vocabulary continuous speech recognition',\n",
       "  'manner of articulation detection',\n",
       "  'speaker identification',\n",
       "  'speaker recognition',\n",
       "  'speaking style synthesis',\n",
       "  'speech recognition',\n",
       "  'speech segmentation',\n",
       "  'speech enhancement',\n",
       "  'speech recognition',\n",
       "  'speech separation',\n",
       "  'speech synthesis',\n",
       "  'speech verification',\n",
       "  'spoken dialogue systems',\n",
       "  'spoken language understanding',\n",
       "  'text independent speaker recognition',\n",
       "  'text to speech synthesis',\n",
       "  'text to speech',\n",
       "  'voice conversion',\n",
       "  'voice query recognition'],\n",
       " 'speech recognition': ['accented speech recognition',\n",
       "  'conversational speech recognition',\n",
       "  'distant speech recognition',\n",
       "  'noisy speech recognition',\n",
       "  'robust speech recognition',\n",
       "  'sequence to sequence speecg recognition',\n",
       "  'visual speech recognition'],\n",
       " 'speech segmentation': ['lexical recognition', 'phonotactic cues'],\n",
       " 'speech emotion recognition': [],\n",
       " 'speech enhancement': [],\n",
       " 'speech separation': [],\n",
       " 'speech synthesis': [],\n",
       " 'speech verification': [],\n",
       " 'spoken dialogue systems': [],\n",
       " 'spoken language understanding': [],\n",
       " 'stacking': [],\n",
       " 'stata': [],\n",
       " 'statistical methods': ['classification methods',\n",
       "  'clustering methods',\n",
       "  'dimensionality reduction methods'],\n",
       " 'statistical based mt': [],\n",
       " 'statistical learning': [],\n",
       " 'steering control': [],\n",
       " 'stemming': ['affix stemmers',\n",
       "  'lemmatisation algorithms',\n",
       "  'matching algorithms',\n",
       "  'stochastic algorithms',\n",
       "  'suffix stripping algorithms',\n",
       "  'n gram analysis'],\n",
       " 'stereo depth estimation': [],\n",
       " 'stereo matching': [],\n",
       " 'stereo from motion': [],\n",
       " 'stochastic gradient descent regression': [],\n",
       " 'stochastic optimization': [],\n",
       " 'stochastic algorithms': [],\n",
       " 'story generation': [],\n",
       " 'style generalization': [],\n",
       " 'style transfer': ['image stylization'],\n",
       " 'sub symbolic based ai': ['behavior based ai',\n",
       "  'nouvelle ai',\n",
       "  'statistical ai',\n",
       "  'computational intelligence',\n",
       "  'embodied intelligence',\n",
       "  'soft computing'],\n",
       " 'suffix stripping algorithms': [],\n",
       " 'super resolution': ['depth map super resolution',\n",
       "  'image super resolution',\n",
       "  'video super resolution',\n",
       "  '3d object super resolution'],\n",
       " 'supervised learning': ['classification',\n",
       "  'ensemble learning',\n",
       "  'non optimization based',\n",
       "  'regression'],\n",
       " 'supervised video summarization': [],\n",
       " 'support vector machine': ['kernel svm', 'radial basis svm'],\n",
       " 'swarm intelligence': [],\n",
       " 'symbolic based ai': ['cognitive simulation',\n",
       "  'knowledge based',\n",
       "  'logic based',\n",
       "  'scruffy'],\n",
       " 'syntax': ['grammar induction',\n",
       "  'lemmatization',\n",
       "  'morphological segmentation',\n",
       "  'parsing',\n",
       "  'pos tagging',\n",
       "  'sbd',\n",
       "  'stemming',\n",
       "  'terminology extraction',\n",
       "  'text segmentaion'],\n",
       " 'synthetic to real translation': [],\n",
       " 'tasks': ['action classification',\n",
       "  'action detection',\n",
       "  'action localization',\n",
       "  'activity recognition',\n",
       "  'animation',\n",
       "  'anomaly detection',\n",
       "  'art analysis',\n",
       "  'audio visual synchronization',\n",
       "  'autonomous vehicles',\n",
       "  'boundary detection',\n",
       "  'color constancy',\n",
       "  'colorization',\n",
       "  'conditional image generation',\n",
       "  'contour detection',\n",
       "  'deblurring',\n",
       "  'deep attention',\n",
       "  'defocus estimation',\n",
       "  'demosaicking',\n",
       "  'denoising',\n",
       "  'dense pixel correspondence estimation',\n",
       "  'depth completion',\n",
       "  'depth estimation',\n",
       "  'disparity estimation',\n",
       "  'document image classification',\n",
       "  'document to image conversion',\n",
       "  'drawing pictures',\n",
       "  'dynamic texture recognition',\n",
       "  'edge detection',\n",
       "  'event based vision',\n",
       "  'face transfer',\n",
       "  'facial recognition and modelling',\n",
       "  'fake image detection',\n",
       "  'few shot learning',\n",
       "  'forgery',\n",
       "  'gaze estimation',\n",
       "  'gesture recognition',\n",
       "  'homography estimation',\n",
       "  'human object interaction detection',\n",
       "  'hyperspectral',\n",
       "  'image to image translation',\n",
       "  'image captioning',\n",
       "  'image classification',\n",
       "  'image clustering',\n",
       "  'image comprehension',\n",
       "  'image compression',\n",
       "  'image cropping',\n",
       "  'image declipping',\n",
       "  'image deconvolution',\n",
       "  'image dehazing',\n",
       "  'image enhancement',\n",
       "  'image generation',\n",
       "  'image imputation',\n",
       "  'image inpainting',\n",
       "  'image manipulation detection',\n",
       "  'image matching',\n",
       "  'image matting',\n",
       "  'image quality assessment',\n",
       "  'image recognition',\n",
       "  'image reconstruction',\n",
       "  'image registration',\n",
       "  'image restoration',\n",
       "  'image retrieval',\n",
       "  'image similarity search',\n",
       "  'image steganography',\n",
       "  'image stitching',\n",
       "  'image categorization',\n",
       "  'infrared and visible image fusion',\n",
       "  'instance search',\n",
       "  'instance segmentation',\n",
       "  'intelligent surveillance',\n",
       "  'interactive segmentation',\n",
       "  'intrinsic image decomposition',\n",
       "  'line segment detection',\n",
       "  'lipreading',\n",
       "  'material classification',\n",
       "  'material recognition',\n",
       "  'metamerism',\n",
       "  'motion capture',\n",
       "  'motion estimation',\n",
       "  'motion segmentation',\n",
       "  'multi modal image segmentation',\n",
       "  'multiview learning',\n",
       "  'music genre transfer',\n",
       "  'novel view synthesis',\n",
       "  'object classification',\n",
       "  'object counting',\n",
       "  'object detection',\n",
       "  'object localization',\n",
       "  'object recognition',\n",
       "  'object reconstruction',\n",
       "  'object tracking',\n",
       "  'observation completion',\n",
       "  'one shot learning',\n",
       "  'optical flow estimation',\n",
       "  'out of distribution detection',\n",
       "  'person re identification',\n",
       "  'person retrieval',\n",
       "  'photo to caricature translation',\n",
       "  'pose estimation',\n",
       "  'saliency detection',\n",
       "  'saliency prediction',\n",
       "  'scene aware dialogue',\n",
       "  'scene classification',\n",
       "  'scene flow estimation',\n",
       "  'scene parsing',\n",
       "  'scene segmentation',\n",
       "  'scene text detection',\n",
       "  'semantic segmentation',\n",
       "  'sensor modeling',\n",
       "  'sign language recognition',\n",
       "  'sign language translation',\n",
       "  'simultaneous localization and mapping',\n",
       "  'sketch recognition',\n",
       "  'spectral estimation',\n",
       "  'stereo matching',\n",
       "  'style generalization',\n",
       "  'style transfer',\n",
       "  'super resolution',\n",
       "  'temporal localization',\n",
       "  'text to image generation',\n",
       "  'text spotting',\n",
       "  'texture classification',\n",
       "  'texture synthesis',\n",
       "  'trajectory prediction',\n",
       "  'unbalanced segmentation',\n",
       "  'vnla',\n",
       "  'video captioning',\n",
       "  'video classification',\n",
       "  'video inpainting',\n",
       "  'video object segmentation',\n",
       "  'video prediction',\n",
       "  'video quality assessment',\n",
       "  'video question answering',\n",
       "  'video summarization',\n",
       "  'viewpoint estimation',\n",
       "  'vision language navigation',\n",
       "  'visual crowd analysis',\n",
       "  'visual localization',\n",
       "  'visual question answering',\n",
       "  'visual sentiment prediction',\n",
       "  'zero shot action recognition',\n",
       "  'zero shot learning',\n",
       "  'image outpainting',\n",
       "  '3d',\n",
       "  'image/document clustering',\n",
       "  'keypoint detection/interest point detection',\n",
       "  'optical character recognition (ocr)'],\n",
       " 'temporal action proposal generation': [],\n",
       " 'temporal difference': [],\n",
       " 'temporal localization': ['camera localization',\n",
       "  'camera relocalization',\n",
       "  'language based temporal localization'],\n",
       " 'tensorflow': [],\n",
       " 'terminology extraction': [],\n",
       " 'text independent speaker recognition': [],\n",
       " 'text to speech synthesis': [],\n",
       " 'text to image generation': [],\n",
       " 'text to speech': ['articulatory synthesis',\n",
       "  'concatenation synthesis',\n",
       "  'deep learning synthesis',\n",
       "  'formant synthesis',\n",
       "  'hmm based synthesis',\n",
       "  'sinewave synthesis'],\n",
       " 'text segmentaion': ['intent segmentation',\n",
       "  'sentence segmentaion',\n",
       "  'topic segmentation',\n",
       "  'word segmentation'],\n",
       " 'text spotting': [],\n",
       " 'textual entailment recognition': [],\n",
       " 'texture': [],\n",
       " 'texture classification': [],\n",
       " 'texture synthesis': [],\n",
       " 'the kalman filter': [],\n",
       " 'theano': [],\n",
       " 'theil sen estimator': [],\n",
       " 'thematics': ['natural language processing', 'perception', 'robotics'],\n",
       " 'thermal infrared object tracking': [],\n",
       " 'thresholding': ['entropy based thresholding',\n",
       "  'global valley approach',\n",
       "  'variance based thresholding',\n",
       "  'adaptive thresholding'],\n",
       " 'time to adjacency analysis': [],\n",
       " 'topic segmentation': [],\n",
       " 'trajectory prediction': [],\n",
       " 'transfer based mt': [],\n",
       " 'transfer reinforcement learning': [],\n",
       " 'tribox method': [],\n",
       " 'umap': [],\n",
       " 'unbalanced segmentation': [],\n",
       " 'uncertain knowledge and reasoning': ['dempster shafer theory',\n",
       "  'fuzzy sets',\n",
       "  'bayesian network',\n",
       "  'decision theory',\n",
       "  'dynamic belief networks',\n",
       "  'dynamic decision networks',\n",
       "  'hidden markov models',\n",
       "  'probabilistic methods for uncertain reasoning',\n",
       "  'utility theory'],\n",
       " 'underwriting': [],\n",
       " 'unit selection synthesis': [],\n",
       " 'unknown environment': [],\n",
       " 'unsupervised anomaly detection': [],\n",
       " 'unsupervised anomaly detection in sound': [],\n",
       " 'unsupervised domain adaptation': [],\n",
       " 'unsupervised image to image translation': ['multimodal unsupervised image to image translation',\n",
       "  'photo to caricature translation',\n",
       "  'synthetic to real translation'],\n",
       " 'unsupervised learning': ['anamoly detection',\n",
       "  'clustering',\n",
       "  'dimensionality reduction'],\n",
       " 'unsupervised object localization': [],\n",
       " 'unsupervised person re identification': [],\n",
       " 'unsupervised representation learning': [],\n",
       " 'unsupervised video object segmentation': [],\n",
       " 'unsupervised video summarization': [],\n",
       " 'utility theory': [],\n",
       " 'vggnet': [],\n",
       " 'vnla': [],\n",
       " 'vanilla gan': [],\n",
       " 'vanilla vae': [],\n",
       " 'variational autoencoder': ['adversarial autoencoder',\n",
       "  'adversarial variational bayes',\n",
       "  'conditional vae',\n",
       "  'denoising vae',\n",
       "  'vanilla vae'],\n",
       " 'video based person re identification': [],\n",
       " 'video captioning': ['audio visual video captioning',\n",
       "  'dense video captioning'],\n",
       " 'video classification': [],\n",
       " 'video inpainting': [],\n",
       " 'video object segmentation': ['semi supervised video object segmentation',\n",
       "  'unsupervised video object segmentation'],\n",
       " 'video prediction': [],\n",
       " 'video quality assessment': [],\n",
       " 'video question answering': [],\n",
       " 'video saliency detection': [],\n",
       " ...}"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "{'absolute error loss': ['mean absolute error (mae)'],\n",
       " 'activation functions': ['arsinh function',\n",
       "  'arctan function',\n",
       "  'bent identity function',\n",
       "  'binary step function',\n",
       "  'bipolar step function',\n",
       "  'elliotsig function',\n",
       "  'gaussian function',\n",
       "  'identity function',\n",
       "  'logistic function',\n",
       "  'maxout function',\n",
       "  'ramp function',\n",
       "  'sq rbf function',\n",
       "  'sqnl function',\n",
       "  'sinc function',\n",
       "  'sinusoid function',\n",
       "  'softexponential function',\n",
       "  'softplus function',\n",
       "  'soft clipping function',\n",
       "  'softmax function',\n",
       "  'tanh function',\n",
       "  'adaptive piecewise linear (apl)',\n",
       "  'bipolar rectified linear unit (brelu)',\n",
       "  'exponential linear unit (elu)',\n",
       "  'gaussian error linear units (gelu)',\n",
       "  'inverse square root linear unit (isrlu)',\n",
       "  'inverse square root unit (isru)',\n",
       "  'leaky rectified linear unit (leaky relu)',\n",
       "  'parameteric rectified linear unit (prelu)',\n",
       "  'randomized leaky rectified linear unit (rrelu)',\n",
       "  'rectified linear unit (relu)',\n",
       "  's shaped rectified linear activation unit (srelu)',\n",
       "  'scaled exponential linear unit (selu)',\n",
       "  'sigmoid linear unit (silu)'],\n",
       " 'adadelta': [],\n",
       " 'adagrad': [],\n",
       " 'adam': [],\n",
       " 'agriculture': [],\n",
       " 'ant colony optimization': [],\n",
       " 'arsinh function': [],\n",
       " 'arctan function': [],\n",
       " 'architecture': [],\n",
       " 'artificial bee colony optimization': [],\n",
       " 'artificial fish swarm algorithm': [],\n",
       " 'artificial intelligence': ['applications', 'fundamental', 'thematics', 'wp'],\n",
       " 'artificial neural networks': ['boltzmann machine',\n",
       "  'hebbian learning',\n",
       "  'hopfield network',\n",
       "  'multilayer feed forward networks',\n",
       "  'perceptrons',\n",
       "  'rnn'],\n",
       " 'arts and entertainment': ['entertainment', 'arts'],\n",
       " 'assistance': ['bots', 'recommendation systems'],\n",
       " 'astronomy': [],\n",
       " 'automobiles': [],\n",
       " 'autonomous vehicles': ['car racing',\n",
       "  'continuous control',\n",
       "  'pedestrian detection',\n",
       "  'steering control'],\n",
       " 'backpropogation': ['levenberg marquardt backpropagation',\n",
       "  'resilient backpropagation (rprop)'],\n",
       " 'bacterial foraging optimization': [],\n",
       " 'banking': [],\n",
       " 'behavior based ai': ['subsumption architecture'],\n",
       " 'bent identity function': [],\n",
       " 'best first search': [],\n",
       " 'bidirectional search': [],\n",
       " 'big data': [],\n",
       " 'binary classification loss functions': ['hinge loss',\n",
       "  'binary cross entropy (bce)'],\n",
       " 'binary sigmoidal function': [],\n",
       " 'binary step function': [],\n",
       " 'bioinformatics': ['dna sequencing', 'genes prediction'],\n",
       " 'biology': ['bioinformatics',\n",
       "  'cell biology',\n",
       "  'ecology',\n",
       "  'fluorescence imaging',\n",
       "  'genetics',\n",
       "  'high throughput imaging',\n",
       "  'microbiology',\n",
       "  'molecular biology',\n",
       "  'neuroscience'],\n",
       " 'bipolar sigmoidal function': [],\n",
       " 'bipolar step function': [],\n",
       " 'blind search': ['brute force search'],\n",
       " 'bots': ['chatbots', 'personal assistant', 'verbal assistant'],\n",
       " 'breadth first search': [],\n",
       " 'brute force search': ['bidirectional search',\n",
       "  'breadth first search',\n",
       "  'depth first search',\n",
       "  'iterative deepening depth first search',\n",
       "  'uniform cost search'],\n",
       " 'business': [],\n",
       " 'cancer': [],\n",
       " 'cell imaging': [],\n",
       " 'cell painting': [],\n",
       " 'cell biology': ['cell imaging', 'cell painting'],\n",
       " 'census': [],\n",
       " 'chatbots': ['conversational bot'],\n",
       " 'classification loss functions': ['binary classification loss functions',\n",
       "  'margin classifier',\n",
       "  'multi class classification loss functions',\n",
       "  'negative log likelihood',\n",
       "  'soft margin classifier'],\n",
       " 'climate': ['weather'],\n",
       " 'computational creativity': [],\n",
       " 'computer security': [],\n",
       " 'construction': [],\n",
       " 'conversational bot': [],\n",
       " 'cosine error': [],\n",
       " 'crime': [],\n",
       " 'cross entropy loss': [],\n",
       " 'culture and humanities': ['culture'],\n",
       " 'cyber security': ['computer security'],\n",
       " 'dna sequencing': [],\n",
       " 'data analytics': [],\n",
       " 'data science': ['data mining'],\n",
       " 'data mining': [],\n",
       " 'demographics': [],\n",
       " 'dempster shafer theory': [],\n",
       " 'depth first search': [],\n",
       " 'design and modelling': [],\n",
       " 'diabetes': [],\n",
       " 'digital preservation': [],\n",
       " 'digital media': [],\n",
       " 'disaster management': ['disaster response',\n",
       "  'forecast system',\n",
       "  'warning systems'],\n",
       " 'distributed artificial intelligence': [],\n",
       " 'ecology': [],\n",
       " 'economics': ['revenue forecast'],\n",
       " 'education': [],\n",
       " 'elliotsig function': ['softsign function'],\n",
       " 'embedding loss functions': ['cosine error', 'l1 hinge error'],\n",
       " 'entertainment': ['movies', 'music', 'photography', 'games'],\n",
       " 'entropy based thresholding': [],\n",
       " 'epidemiology': [],\n",
       " 'evolutionary computation': [],\n",
       " 'feature engineering': [],\n",
       " 'film': [],\n",
       " 'finance': [],\n",
       " 'first order optimization algorithms': ['adadelta',\n",
       "  'adagrad',\n",
       "  'adam',\n",
       "  'gradient descent',\n",
       "  'mini batch gradient descent',\n",
       "  'nesterov accelerated gradient',\n",
       "  'stochastic gradient descent'],\n",
       " 'flower pollination optimization': [],\n",
       " 'fluorescence imaging': [],\n",
       " 'fuzzy logic': [],\n",
       " 'fuzzy sets': ['fuzzy logic'],\n",
       " 'fuzzy systems': [],\n",
       " 'gru': [],\n",
       " 'gaussian function': [],\n",
       " 'genetic algorithm optimization': [],\n",
       " 'genetics': ['genomics'],\n",
       " 'genomics': [],\n",
       " 'geography': [],\n",
       " 'geospatial': [],\n",
       " 'global valley approach': [],\n",
       " 'gradient descent': [],\n",
       " 'gravitational search algorithm': [],\n",
       " 'greedy search': ['best first search'],\n",
       " 'health': ['health care'],\n",
       " 'heuristic search': ['greedy search',\n",
       "  'pure heuristic search',\n",
       "  'a* search',\n",
       "  'iterative deepening a* search'],\n",
       " 'high throughput imaging': [],\n",
       " 'hill climbing search': [],\n",
       " 'hinge loss': [],\n",
       " 'housing': [],\n",
       " 'huber loss': [],\n",
       " 'human computer interaction': [],\n",
       " 'identity function': [],\n",
       " 'image processing': ['digital image processing'],\n",
       " 'income': ['tax'],\n",
       " 'infrastructure': [],\n",
       " 'intelligent water drops algorithm': [],\n",
       " 'intrusion detection': [],\n",
       " 'iterative deepening depth first search': [],\n",
       " 'journalism': ['news'],\n",
       " 'kullback leibler divergence loss': [],\n",
       " 'l1 hinge error': [],\n",
       " 'l1 loss': [],\n",
       " 'l2 loss': [],\n",
       " 'law': [],\n",
       " 'levenberg marquardt backpropagation': [],\n",
       " 'life sciences': [],\n",
       " 'literature': [],\n",
       " 'local beam search': [],\n",
       " 'local search': ['hill climbing search',\n",
       "  'local beam search',\n",
       "  'simulated annealing',\n",
       "  'travelling salesman problem'],\n",
       " 'logic based reasoning': ['first order logic',\n",
       "  'logical reasoning',\n",
       "  'modal logic',\n",
       "  'non monotonic logic',\n",
       "  'proportional logic'],\n",
       " 'logistic function': ['sigmoid function', 'softstep function'],\n",
       " 'manufacturing': [],\n",
       " 'margin classifier': [],\n",
       " 'marketing': [],\n",
       " 'mathematics': ['probability', 'statistics'],\n",
       " 'maxout function': [],\n",
       " 'mean bias error': [],\n",
       " 'media': ['digital media', 'journalism', 'multimedia'],\n",
       " 'medicine': ['epidemiology'],\n",
       " 'meteorology': ['climate', 'geographic information system (gis)'],\n",
       " 'microbiology': [],\n",
       " 'migrating birds optimization': [],\n",
       " 'mini max algorithm': [],\n",
       " 'mini batch gradient descent': [],\n",
       " 'miscellaneous ai applications': ['agriculture',\n",
       "  'assistance',\n",
       "  'astronomy',\n",
       "  'big data',\n",
       "  'crime',\n",
       "  'culture and humanities',\n",
       "  'data analytics',\n",
       "  'digital preservation',\n",
       "  'disaster management',\n",
       "  'law',\n",
       "  'literature',\n",
       "  'media',\n",
       "  'meteorology',\n",
       "  'nature and life sciences',\n",
       "  'politics',\n",
       "  'security',\n",
       "  'sociology',\n",
       "  'survey',\n",
       "  'urban development',\n",
       "  'imaging',\n",
       "  'satellite imaging/satellite imagery'],\n",
       " 'molecular docking': [],\n",
       " 'molecular biology': ['molecular docking'],\n",
       " 'movies': ['film'],\n",
       " 'multi class classification loss functions': ['kullback leibler divergence loss',\n",
       "  'multi class cross entropy loss'],\n",
       " 'multi class cross entropy loss': [],\n",
       " 'multiclass classification': [],\n",
       " 'music': [],\n",
       " 'nature and life sciences': ['biology',\n",
       "  'life sciences',\n",
       "  'environmental',\n",
       "  'microscopy'],\n",
       " 'negative log likelihood': ['cross entropy loss'],\n",
       " 'nesterov accelerated gradient': [],\n",
       " 'news': [],\n",
       " 'nouvelle ai': [],\n",
       " 'optical imagery': [],\n",
       " 'optimization algorithms': ['ant colony optimization',\n",
       "  'artificial bee colony optimization',\n",
       "  'artificial fish swarm algorithm',\n",
       "  'bacterial foraging optimization',\n",
       "  'first order optimization algorithms',\n",
       "  'flower pollination optimization',\n",
       "  'genetic algorithm optimization',\n",
       "  'intelligent water drops algorithm',\n",
       "  'migrating birds optimization',\n",
       "  'mini max algorithm',\n",
       "  'particle swarm optimization',\n",
       "  'second order optimization algorithms'],\n",
       " 'particle swarm optimization': [],\n",
       " 'performance metrics': ['accuracy metrics'],\n",
       " 'personal assistant': [],\n",
       " 'pharmaceutical': [],\n",
       " 'photography': [],\n",
       " 'politics': [],\n",
       " 'probability': [],\n",
       " 'pure heuristic search': [],\n",
       " 'quadratic loss': [],\n",
       " 'ramp function': [],\n",
       " 'recommendation systems': ['collaborative filtering',\n",
       "  'content based filtering',\n",
       "  'session based recommendations'],\n",
       " 'regression loss functions': ['absolute error loss',\n",
       "  'huber loss',\n",
       "  'mean bias error',\n",
       "  'squared error loss'],\n",
       " 'remote sensing': [],\n",
       " 'revenue forecast': [],\n",
       " 'sq rbf function': [],\n",
       " 'sqnl function': [],\n",
       " 'search algorithms': ['blind search',\n",
       "  'gravitational search algorithm',\n",
       "  'heuristic search',\n",
       "  'local search',\n",
       "  'tabu search algorithm'],\n",
       " 'second order optimization algorithms': [],\n",
       " 'security': ['cyber security', 'intrusion detection', 'surveillance'],\n",
       " 'sigmoid function': ['sigmoidal function'],\n",
       " 'sigmoidal function': ['binary sigmoidal function',\n",
       "  'bipolar sigmoidal function'],\n",
       " 'simulated annealing': [],\n",
       " 'sinc function': [],\n",
       " 'sinusoid function': [],\n",
       " 'sociology': [],\n",
       " 'softexponential function': [],\n",
       " 'softplus function': [],\n",
       " 'soft clipping function': [],\n",
       " 'soft margin classifier': [],\n",
       " 'softmax function': [],\n",
       " 'softsign function': [],\n",
       " 'softstep function': [],\n",
       " 'speech denoising': [],\n",
       " 'speech processing': [],\n",
       " 'sports': [],\n",
       " 'squared error loss': ['mean square error (mse)'],\n",
       " 'statistical ai': [],\n",
       " 'statistics': [],\n",
       " 'stochastic gradient descent': [],\n",
       " 'subsumption architecture': [],\n",
       " 'surveillance': [],\n",
       " 'survey': ['census'],\n",
       " 'tabu search algorithm': [],\n",
       " 'tanh function': [],\n",
       " 'tax': [],\n",
       " 'terrain mapping': [],\n",
       " 'text mining': [],\n",
       " 'trade': [],\n",
       " 'traffic': [],\n",
       " 'transfer learning': [],\n",
       " 'transport': [],\n",
       " 'transportation': ['automobiles',\n",
       "  'autonomous vehicles',\n",
       "  'traffic',\n",
       "  'transport'],\n",
       " 'travelling salesman problem': [],\n",
       " 'uniform cost search': [],\n",
       " 'urban development': ['architecture',\n",
       "  'banking',\n",
       "  'construction',\n",
       "  'demographics',\n",
       "  'design and modelling',\n",
       "  'education',\n",
       "  'housing',\n",
       "  'income',\n",
       "  'infrastructure',\n",
       "  'transportation'],\n",
       " 'variance based thresholding': [],\n",
       " 'verbal assistant': [],\n",
       " 'video games': [],\n",
       " 'visual arts': [],\n",
       " 'weather': [],\n",
       " 'aerial imagery': [],\n",
       " 'culture': [],\n",
       " 'disaster response': [],\n",
       " 'emissions': [],\n",
       " 'environmental': ['geography',\n",
       "  'emissions',\n",
       "  'pollution',\n",
       "  'quality',\n",
       "  'radiations'],\n",
       " 'forecast system': [],\n",
       " 'genes prediction': [],\n",
       " 'imaging': [],\n",
       " 'microscopy': [],\n",
       " 'multimedia': [],\n",
       " 'neuro biology': [],\n",
       " 'neuro imaging': [],\n",
       " 'neuroscience': ['neuro biology', 'neuro imaging'],\n",
       " 'pollution': [],\n",
       " 'quality': [],\n",
       " 'radiations': [],\n",
       " 'warning systems': [],\n",
       " 'a* search': [],\n",
       " 'adaptive piecewise linear (apl)': [],\n",
       " 'binary cross entropy (bce)': [],\n",
       " 'bipolar rectified linear unit (brelu)': [],\n",
       " 'exponential linear unit (elu)': [],\n",
       " 'gaussian error linear units (gelu)': [],\n",
       " 'geographic information system (gis)': [],\n",
       " 'geoinformatics (gis)': [],\n",
       " 'inverse square root linear unit (isrlu)': [],\n",
       " 'inverse square root unit (isru)': [],\n",
       " 'iterative deepening a* search': [],\n",
       " 'leaky rectified linear unit (leaky relu)': [],\n",
       " 'mean absolute error (mae)': ['l1 loss'],\n",
       " 'mean square error (mse)': ['l2 loss', 'quadratic loss'],\n",
       " 'parameteric rectified linear unit (prelu)': [],\n",
       " 'randomized leaky rectified linear unit (rrelu)': [],\n",
       " 'rectified linear unit (relu)': [],\n",
       " 'resilient backpropagation (rprop)': [],\n",
       " 's shaped rectified linear activation unit (srelu)': [],\n",
       " 'satellite imaging/satellite imagery': ['geospatial',\n",
       "  'optical imagery',\n",
       "  'remote sensing',\n",
       "  'terrain mapping',\n",
       "  'aerial imagery',\n",
       "  'geoinformatics (gis)'],\n",
       " 'scaled exponential linear unit (selu)': [],\n",
       " 'sigmoid linear unit (silu)': [],\n",
       " 'ahe': [],\n",
       " 'aiml': [],\n",
       " 'ai approaches': ['cybernetics and brain simulation',\n",
       "  'statistical learning',\n",
       "  'sub symbolic based ai',\n",
       "  'symbolic based ai'],\n",
       " 'ai concepts': ['ai software library',\n",
       "  'affective computing',\n",
       "  'evolutionary algorithms',\n",
       "  'game theory',\n",
       "  'human computer interaction',\n",
       "  'knowledge engineering',\n",
       "  'ontology',\n",
       "  'sequential learning',\n",
       "  'swarm intelligence',\n",
       "  'transfer learning'],\n",
       " 'ai methodoogy': ['activation functions',\n",
       "  'backpropogation',\n",
       "  'performance metrics',\n",
       "  'active learning',\n",
       "  'automl',\n",
       "  'bayesian inference',\n",
       "  'combinatorial optimization',\n",
       "  'data augmentation',\n",
       "  'density estimation',\n",
       "  'distributional reinforcement learning',\n",
       "  'domain adaptation',\n",
       "  'entity embeddings',\n",
       "  'feature importance',\n",
       "  'feature selection',\n",
       "  'hierarchical reinforcement learning',\n",
       "  'imitation learning',\n",
       "  'interpretable machine learning',\n",
       "  'loss function',\n",
       "  'meta learning',\n",
       "  'metric learning',\n",
       "  'model compression',\n",
       "  'model selection',\n",
       "  'multi goal reinforcement learning',\n",
       "  'multi agent reinforcement learning',\n",
       "  'multiple instance learning',\n",
       "  'network pruning',\n",
       "  'outlier detection',\n",
       "  'policy gradient methods',\n",
       "  'quantization',\n",
       "  'recommendation systems',\n",
       "  'representation learning',\n",
       "  'sparse learning',\n",
       "  'stochastic optimization',\n",
       "  'transfer learning'],\n",
       " 'ai ontology': ['artificial intelligence'],\n",
       " 'ai programming language': ['aiml',\n",
       "  'cplusplus',\n",
       "  'c language',\n",
       "  'centura',\n",
       "  'haskell',\n",
       "  'java',\n",
       "  'julia',\n",
       "  'lisp',\n",
       "  'matlab',\n",
       "  'perl',\n",
       "  'prolog',\n",
       "  'python',\n",
       "  'r language',\n",
       "  'ruby'],\n",
       " 'ai software library': ['ai programming language',\n",
       "  'azure',\n",
       "  'accord.net',\n",
       "  'apachemahout',\n",
       "  'caffe',\n",
       "  'dllearner',\n",
       "  'elki',\n",
       "  'eviews',\n",
       "  'encog',\n",
       "  'fama',\n",
       "  'h2o',\n",
       "  'ibmminer',\n",
       "  'knime',\n",
       "  'kxen',\n",
       "  'keras',\n",
       "  'lionsolver',\n",
       "  'liblinear',\n",
       "  'libsvm',\n",
       "  'mlpack',\n",
       "  'moa',\n",
       "  'mx net',\n",
       "  'mathematica',\n",
       "  'microsoft cntk',\n",
       "  'montecarlomachinelearninglibrary',\n",
       "  'neurosolutions',\n",
       "  'octave',\n",
       "  'opencv',\n",
       "  'opennn',\n",
       "  'oracledatamining',\n",
       "  'orange',\n",
       "  'pytorch',\n",
       "  'rcase',\n",
       "  'rapidminer',\n",
       "  'sap',\n",
       "  'sas',\n",
       "  'spss',\n",
       "  'sqlserveranalysisservices',\n",
       "  'statistica',\n",
       "  'shogun',\n",
       "  'spark mlib',\n",
       "  'stata',\n",
       "  'tensorflow',\n",
       "  'theano',\n",
       "  'weka',\n",
       "  'yale',\n",
       "  'yooreeka',\n",
       "  'mlpy',\n",
       "  'scikit learn'],\n",
       " 'atc': [],\n",
       " 'azure': [],\n",
       " 'accented speech recognition': [],\n",
       " 'accord.net': [],\n",
       " 'accuracy metrics': [],\n",
       " 'acoustic modelling': [],\n",
       " 'acoustic question answering': [],\n",
       " 'action classification': [],\n",
       " 'action detection': ['fine grained action detection',\n",
       "  'multiple action detection',\n",
       "  'skeleton based action recognition',\n",
       "  'temporal action proposal generation'],\n",
       " 'action localization': ['spatio temporal action localization'],\n",
       " 'active learning': ['known environment', 'unknown environment'],\n",
       " 'active object localization': [],\n",
       " 'activity recognition': ['egocentric activity recognition',\n",
       "  'group activity recognition',\n",
       "  'human activity recognition'],\n",
       " 'ad targeting': [],\n",
       " 'adaboost': [],\n",
       " 'adaptive thresholding': [],\n",
       " 'adversarial autoencoder': [],\n",
       " 'adversarial variational bayes': [],\n",
       " 'affective computing': [],\n",
       " 'affinity propogation': [],\n",
       " 'affix stemmers': [],\n",
       " 'agents and environment': [],\n",
       " 'agglomerative': [],\n",
       " 'alexnet': [],\n",
       " 'algorithmic trading': [],\n",
       " 'analogy': [],\n",
       " 'anamoly detection': ['elliptic envelope',\n",
       "  'local outlier factor',\n",
       "  'one class svm'],\n",
       " 'animation': [],\n",
       " 'anomaly detection': ['anomaly detection in surveillance videos',\n",
       "  'group anomaly detection',\n",
       "  'unsupervised anomaly detection',\n",
       "  'unsupervised anomaly detection in sound'],\n",
       " 'anomaly detection in surveillance videos': [],\n",
       " 'apachemahout': [],\n",
       " 'applications': ['arts and entertainment',\n",
       "  'health',\n",
       "  'miscellaneous ai applications',\n",
       "  'audit',\n",
       "  'automotive',\n",
       "  'finance and economics',\n",
       "  'marketing and advertising',\n",
       "  'military'],\n",
       " 'art analysis': [],\n",
       " 'articulatory synthesis': [],\n",
       " 'arts': ['linguistics', 'musical', 'visual and artistic', 'problem solving'],\n",
       " 'audio visual speech recognition': [],\n",
       " 'audio visual synchronization': [],\n",
       " 'audio visual video captioning': [],\n",
       " 'audit': [],\n",
       " 'automl': ['automated feature engineering',\n",
       "  'hyperparameter optimization',\n",
       "  'neural architecture search'],\n",
       " 'autoencoder': ['variational autoencoder'],\n",
       " 'autoencoders': [],\n",
       " 'automated feature engineering': [],\n",
       " 'automatic relevance determination': [],\n",
       " 'automatic summarization': [],\n",
       " 'automotive': ['manufacturing', 'transportation'],\n",
       " 'auxiliary classifier gan': [],\n",
       " 'auxiliary learning': [],\n",
       " 'backward feature elimination': [],\n",
       " 'basic image filtering': [],\n",
       " 'bayesian inference': ['probabilistic programming'],\n",
       " 'bayesian network': [],\n",
       " 'bayesian regression': ['automatic relevance determination',\n",
       "  'bayesian ridge regression'],\n",
       " 'bayesian ridge regression': [],\n",
       " 'belief network': [],\n",
       " 'bidirectional rnn': [],\n",
       " 'binary rbm with contrastive divergence': [],\n",
       " 'binary rbm with persistent contrastive divergence': [],\n",
       " 'binary shape analysis': ['boundary tracking',\n",
       "  'object labeling and counting',\n",
       "  'shape recognition',\n",
       "  'size filtering',\n",
       "  'skeletons and thinning'],\n",
       " 'birch': [],\n",
       " 'bird view synthesis': [],\n",
       " 'blind image deblurring': [],\n",
       " 'blind image quality assessment': [],\n",
       " 'boltzmann machine': ['dbm', 'rbm'],\n",
       " 'boosting': ['adaboost',\n",
       "  'brownboost',\n",
       "  'graddient boosting',\n",
       "  'lpboost',\n",
       "  'logitboost'],\n",
       " 'boundary detection': [],\n",
       " 'boundary equilibrium gan': [],\n",
       " 'boundary pattern analysis': [],\n",
       " 'boundary seeking gan': [],\n",
       " 'boundary tracking': [],\n",
       " 'brill tagger': [],\n",
       " 'brownboost': [],\n",
       " 'byte pair encoding': [],\n",
       " 'clahe': [],\n",
       " 'cnn': ['alexnet',\n",
       "  'lenet',\n",
       "  'mask rcnn',\n",
       "  'rcnn',\n",
       "  'resnet',\n",
       "  'vggnet',\n",
       "  'zfnet',\n",
       "  'googlenet/inception'],\n",
       " 'cplusplus': [],\n",
       " 'c language': [],\n",
       " 'caffe': [],\n",
       " 'camera localization': [],\n",
       " 'camera relocalization': [],\n",
       " 'car racing': [],\n",
       " 'catboost': [],\n",
       " 'centura': [],\n",
       " 'circle and ellipse detection': [],\n",
       " 'classification': ['multiclass classification',\n",
       "  'decision trees',\n",
       "  'gradient boost',\n",
       "  'logistic regression',\n",
       "  'naive bayes classifier',\n",
       "  'nearest neighbor',\n",
       "  'random forest',\n",
       "  'support vector machine'],\n",
       " 'classification methods': [],\n",
       " 'clinical training': [],\n",
       " 'clustering': ['affinity propogation',\n",
       "  'birch',\n",
       "  'density based spatial clustering',\n",
       "  'expectation maximization',\n",
       "  'hierarchical clustering',\n",
       "  'k means clustering',\n",
       "  'mean shift clustering',\n",
       "  'mixture model',\n",
       "  'optics algorithm',\n",
       "  'self organizing maps',\n",
       "  'spectral clustering'],\n",
       " 'clustering methods': [],\n",
       " 'co saliency detection': [],\n",
       " 'cognitive simulation': [],\n",
       " 'collaborative filtering': [],\n",
       " 'color constancy': ['few shot camera adaptive color constancy'],\n",
       " 'color image compression artifact reduction': [],\n",
       " 'colorization': [],\n",
       " 'combinatorial optimization': [],\n",
       " 'computational intelligence': [],\n",
       " 'computer aided interpretation of medical images': [],\n",
       " 'computer vision': ['image processing',\n",
       "  'intermediate level vision',\n",
       "  'low level vision',\n",
       "  'machine vision',\n",
       "  'neural network based',\n",
       "  'pattern recognition',\n",
       "  'tasks',\n",
       "  '3d vision'],\n",
       " 'concatenation synthesis': ['diphone synthesis',\n",
       "  'domain specific synthesis',\n",
       "  'unit selection synthesis'],\n",
       " 'conditional image generation': [],\n",
       " 'conditional planning': [],\n",
       " 'conditional vae': [],\n",
       " 'conditional random fields approach': [],\n",
       " 'constrained diffeomorphic image registration': [],\n",
       " 'content based filtering': [],\n",
       " 'content based image retrieval': [],\n",
       " 'content generation': [],\n",
       " 'context based emotion recognition': [],\n",
       " 'continuous control': [],\n",
       " 'continuous object recognition': [],\n",
       " 'contour detection': [],\n",
       " 'conversational speech recognition': [],\n",
       " 'converssational emotion recognition': [],\n",
       " 'convolutional neural networks': [],\n",
       " 'coreference resolution': [],\n",
       " 'corner and interest point detection': ['local invariant feature detectors and descriptors'],\n",
       " 'coupled gan': [],\n",
       " 'curved text detection': [],\n",
       " 'cybernetics and brain simulation': [],\n",
       " 'dbm': [],\n",
       " 'dllearner': [],\n",
       " 'data augmentation': [],\n",
       " 'deblurring': ['blind image deblurring', 'single image blind deblurring'],\n",
       " 'decision tree regressor': [],\n",
       " 'decision stump': [],\n",
       " 'decision theory': [],\n",
       " 'decision trees': ['decision stump'],\n",
       " 'deep adversarial networks': [],\n",
       " 'deep attention': [],\n",
       " 'deep learning': ['transfer learning',\n",
       "  'convolutional neural networks',\n",
       "  'recurrent neural networks',\n",
       "  'recursive neural networks',\n",
       "  'unsupervised pretrained networks (upn)'],\n",
       " 'deep rnn': [],\n",
       " 'deep reinforcement learning': [],\n",
       " 'deep feedforward and recurrent neural networks': [],\n",
       " 'deep learning synthesis': [],\n",
       " 'default reasoning and the qualification problem': [],\n",
       " 'defocus estimation': [],\n",
       " 'deformable object manipulation': [],\n",
       " 'demosaicking': [],\n",
       " 'denoising': ['speech denoising', 'image denoising'],\n",
       " 'denoising vae': [],\n",
       " 'dense pixel correspondence estimation': [],\n",
       " 'dense video captioning': [],\n",
       " 'density based spatial clustering': [],\n",
       " 'density estimation': [],\n",
       " 'depiction invariant object recognition': [],\n",
       " 'depth and camera motion': [],\n",
       " 'depth completion': [],\n",
       " 'depth estimation': ['depth and camera motion',\n",
       "  'depth prediction for outdoor scene',\n",
       "  'monocular depth estimation',\n",
       "  'stereo depth estimation'],\n",
       " 'depth map super resolution': [],\n",
       " 'depth prediction for outdoor scene': [],\n",
       " 'description logic': [],\n",
       " 'design treatment plans': [],\n",
       " 'dictionary based mt': [],\n",
       " 'digit recognition': [],\n",
       " 'digital image processing': ['ahe', 'clahe', 'gcn'],\n",
       " 'dimensionality reduction': ['autoencoder',\n",
       "  'backward feature elimination',\n",
       "  'factor analysis',\n",
       "  'forward feature selection',\n",
       "  'high correlation filter',\n",
       "  'ica',\n",
       "  'low variance filter',\n",
       "  'methods based on projections',\n",
       "  'missing value ratio',\n",
       "  'pca',\n",
       "  'projection pursuit',\n",
       "  'svd',\n",
       "  'sammon mapping',\n",
       "  'umap',\n",
       "  't sne',\n",
       "  'flexible discriminant analysis (fda)',\n",
       "  'mixture discriminant analysis (mda)',\n",
       "  'multidimensional scaling (mds)',\n",
       "  'partial least squares regression (plsr)',\n",
       "  'principal component regression (pcr)',\n",
       "  'quadratic discriminant analysis (qda)'],\n",
       " 'dimensionality reduction methods': [],\n",
       " 'diphone synthesis': [],\n",
       " 'discogan': [],\n",
       " 'discourse': ['automatic summarization', 'coreference resolution'],\n",
       " 'disease prediction and diagnosis': ['cancer', 'diabetes'],\n",
       " 'disparity estimation': [],\n",
       " 'distant speech recognition': [],\n",
       " 'distributional reinforcement learning': [],\n",
       " 'distributional semantics': ['hyperspace analogue to language',\n",
       "  'latent semantic analysis',\n",
       "  'random indexing',\n",
       "  'semantic folding'],\n",
       " 'distributional learning': [],\n",
       " 'divisive': [],\n",
       " 'document embedding': [],\n",
       " 'document image classification': [],\n",
       " 'document to image conversion': [],\n",
       " 'domain specific synthesis': [],\n",
       " 'domain adaptation': ['domain generalization',\n",
       "  'partial domain adaptation',\n",
       "  'predictive domain adaptation',\n",
       "  'unsupervised domain adaptation'],\n",
       " 'domain generalization': [],\n",
       " 'drawing pictures': [],\n",
       " 'drug creation': ['molecular docking', 'pharmaceutical'],\n",
       " 'dualgan': [],\n",
       " 'dynamic belief networks': [],\n",
       " 'dynamic decision networks': [],\n",
       " 'dynamic pricing': [],\n",
       " 'dynamic texture recognition': [],\n",
       " 'dynamic time warping': [],\n",
       " 'elki': [],\n",
       " 'eviews': [],\n",
       " 'edge detection': [],\n",
       " 'effector': [],\n",
       " 'egocentric activity recognition': [],\n",
       " 'elastic net': [],\n",
       " 'elliptic envelope': [],\n",
       " 'embodied question answering': [],\n",
       " 'embodied intelligence': [],\n",
       " 'emotion recognition': ['context based emotion recognition',\n",
       "  'converssational emotion recognition',\n",
       "  'multimodal emotion recognition',\n",
       "  'speech emotion recognition'],\n",
       " 'encog': [],\n",
       " 'energy based gan': [],\n",
       " 'ensemble learning': ['boosting',\n",
       "  'random forest',\n",
       "  'stacking',\n",
       "  'bootstrap aggregating (bagging)'],\n",
       " 'entity embeddings': [],\n",
       " 'event based vision': [],\n",
       " 'evolutionary algorithms': ['genetic algorithms'],\n",
       " 'example based mt': [],\n",
       " 'expectation maximization': [],\n",
       " 'expressive operator descriptions': [],\n",
       " 'fama': [],\n",
       " 'face alignment': [],\n",
       " 'face detection': [],\n",
       " 'face generation': [],\n",
       " 'face recognition': [],\n",
       " 'face reconstruction': [],\n",
       " 'face transfer': [],\n",
       " 'face verification': [],\n",
       " 'facial expression recognition': [],\n",
       " 'facial inpainting': [],\n",
       " 'facial recognition and modelling': ['face alignment',\n",
       "  'face detection',\n",
       "  'face generation',\n",
       "  'face recognition',\n",
       "  'face reconstruction',\n",
       "  'face verification',\n",
       "  'facial expression recognition'],\n",
       " 'factor analysis': [],\n",
       " 'factual visual question answering': [],\n",
       " 'fake image detection': [],\n",
       " 'feature importance': [],\n",
       " 'feature learning': ['deep learning', 'deep reinforcement learning'],\n",
       " 'feature selection': [],\n",
       " 'few shot camera adaptive color constancy': [],\n",
       " 'few shot image classification': [],\n",
       " 'few shot imitation learning': [],\n",
       " 'few shot learning': ['few shot camera adaptive color constancy',\n",
       "  'few shot imitation learning',\n",
       "  'few shot relation classification'],\n",
       " 'few shot relation classification': [],\n",
       " 'finance and economics': ['business',\n",
       "  'economics',\n",
       "  'finance',\n",
       "  'algorithmic trading',\n",
       "  'fraud detection',\n",
       "  'market analysis and data mining',\n",
       "  'personal finance',\n",
       "  'portfolio management',\n",
       "  'underwriting'],\n",
       " 'fine grained action detection': [],\n",
       " 'fine grained image classification': [],\n",
       " 'fine grained image recognition': [],\n",
       " 'fine grain image categorization': [],\n",
       " 'first order logic': [],\n",
       " 'forgery': [],\n",
       " 'formant synthesis': [],\n",
       " 'forward feature selection': [],\n",
       " 'frame systems and semantic networks': [],\n",
       " 'fraud detection': [],\n",
       " 'fraudalent ads': [],\n",
       " 'fundamental': ['distributed artificial intelligence',\n",
       "  'mathematics',\n",
       "  'ai approaches',\n",
       "  'ai concepts',\n",
       "  'ai methodoogy',\n",
       "  'general intelligence',\n",
       "  'knowledge and reasoning',\n",
       "  'machine learning',\n",
       "  'planning',\n",
       "  'problem solving',\n",
       "  'search and optimization',\n",
       "  'social intelligence',\n",
       "  'uncertain knowledge and reasoning'],\n",
       " 'gan': ['auxiliary classifier gan',\n",
       "  'boundary equilibrium gan',\n",
       "  'boundary seeking gan',\n",
       "  'coupled gan',\n",
       "  'discogan',\n",
       "  'dualgan',\n",
       "  'energy based gan',\n",
       "  'generative adversarial parallelization',\n",
       "  'gibbsnet',\n",
       "  'improved training for wasserstein gan',\n",
       "  'infogan',\n",
       "  'least squares gan',\n",
       "  'magan',\n",
       "  'mode regularized gan',\n",
       "  'softmax gan',\n",
       "  'vanilla gan',\n",
       "  'wasserstein gan',\n",
       "  'f gan',\n",
       "  'adversarial feature learning & adversarially learned inference',\n",
       "  'conditional gan (cgan)',\n",
       "  'deep convolutional gan (dcgan)',\n",
       "  'laplacian pyramid gan (lapgan)',\n",
       "  'super resolution gan (srgan)'],\n",
       " 'gbm': [],\n",
       " 'gcn': [],\n",
       " 'game theory': [],\n",
       " 'games': ['sports', 'video games'],\n",
       " 'gated recurrent unit': ['gru'],\n",
       " 'gaze estimation': [],\n",
       " 'general intelligence': ['agents and environment'],\n",
       " 'general logical descriptions': [],\n",
       " 'general ontology': [],\n",
       " 'generalized zero shot learning': [],\n",
       " 'generative adversarial networks': ['gan'],\n",
       " 'generative adversarial parallelization': [],\n",
       " 'genetic algorithms': [],\n",
       " 'gesture recognition': ['hand gesture recognition',\n",
       "  'isolated gestures recognition'],\n",
       " 'gibbsnet': [],\n",
       " 'graddient boosting': [],\n",
       " 'gradient boost regressor': [],\n",
       " 'gradient boost': ['catboost',\n",
       "  'catboost',\n",
       "  'gbm',\n",
       "  'lightgbm',\n",
       "  'light gbm',\n",
       "  'xgboost',\n",
       "  'xgboost'],\n",
       " 'grammar induction': ['distributional learning',\n",
       "  'grammatical inference by genetic algorithms',\n",
       "  'grammatical inference by greedy algorithms',\n",
       "  'grammatical inference by trial and error',\n",
       "  'pattern theory'],\n",
       " 'grammatical inference by genetic algorithms': [],\n",
       " 'grammatical inference by greedy algorithms': ['byte pair encoding',\n",
       "  'lempel ziv welch algorithm',\n",
       "  'sequitur'],\n",
       " 'grammatical inference by trial and error': [],\n",
       " 'graph embedding': [],\n",
       " 'graph representation learning': [],\n",
       " 'group activity recognition': [],\n",
       " 'group anomaly detection': [],\n",
       " 'h2o': [],\n",
       " 'hmm based synthesis': [],\n",
       " 'hand gesture recognition': [],\n",
       " 'handwritten digit recognition': [],\n",
       " 'haskell': [],\n",
       " 'health care': ['medicine',\n",
       "  'clinical training',\n",
       "  'computer aided interpretation of medical images',\n",
       "  'design treatment plans',\n",
       "  'disease prediction and diagnosis',\n",
       "  'drug creation',\n",
       "  'medication management',\n",
       "  'provide consultations and companion robots'],\n",
       " 'hebbian learning': [],\n",
       " 'hidden markov models based': [],\n",
       " 'hidden markov models': [],\n",
       " 'hierarchical clustering': ['agglomerative', 'divisive'],\n",
       " 'hierarchical decomposition': [],\n",
       " 'hierarchical reinforcement learning': [],\n",
       " 'high correlation filter': [],\n",
       " 'homography estimation': [],\n",
       " 'hopfield network': [],\n",
       " 'hough transform': [],\n",
       " 'huber regression': [],\n",
       " 'human object interaction detection': ['human computer interaction'],\n",
       " 'human activity recognition': [],\n",
       " 'hybrid mt': [],\n",
       " 'hyperparameter optimization': [],\n",
       " 'hyperspace analogue to language': [],\n",
       " 'hyperspectral': ['hyperspectral image segmentation',\n",
       "  'hyperspectral unmixing'],\n",
       " 'hyperspectral image classification': [],\n",
       " 'hyperspectral image segmentation': [],\n",
       " 'hyperspectral unmixing': [],\n",
       " 'hypothesis of creative patterns': [],\n",
       " 'ibmminer': [],\n",
       " 'ica': [],\n",
       " 'image based localization': [],\n",
       " 'image to image translation': ['bird view synthesis',\n",
       "  'keypoint localization',\n",
       "  'optical character recognition',\n",
       "  'unsupervised image to image translation'],\n",
       " 'image captioning': [],\n",
       " 'image classification': ['few shot image classification',\n",
       "  'fine grained image classification',\n",
       "  'hyperspectral image classification',\n",
       "  'semi supervised image classification'],\n",
       " 'image clustering': [],\n",
       " 'image comprehension': [],\n",
       " 'image compression': ['color image compression artifact reduction',\n",
       "  'jpeg compression artifact reduction',\n",
       "  'lossy compression artifact reduction'],\n",
       " 'image cropping': [],\n",
       " 'image declipping': [],\n",
       " 'image deconvolution': [],\n",
       " 'image dehazing': [],\n",
       " 'image denoising': ['salt and pepper noise removal'],\n",
       " 'image enhancement': ['local color enhancement',\n",
       "  'low light image enhancement'],\n",
       " 'image generation': [],\n",
       " 'image imputation': [],\n",
       " 'image inpainting': ['facial inpainting'],\n",
       " 'image manipulation detection': [],\n",
       " 'image matching': [],\n",
       " 'image matting': [],\n",
       " 'image quality assessment': ['blind image quality assessment',\n",
       "  'no reference image quality assessment'],\n",
       " 'image recognition': ['fine grained image recognition'],\n",
       " 'image reconstruction': [],\n",
       " 'image rectification': [],\n",
       " 'image registration': ['constrained diffeomorphic image registration'],\n",
       " 'image restoration': [],\n",
       " 'image retrieval': ['content based image retrieval',\n",
       "  'multi label image retrieval',\n",
       "  'sketch based image retrieval'],\n",
       " 'image similarity search': [],\n",
       " 'image steganography': [],\n",
       " 'image stitching': [],\n",
       " 'image stylization': [],\n",
       " 'image super resolution': [],\n",
       " 'image transformations and camera calibration': ['image rectification',\n",
       "  '3 d reconstruction'],\n",
       " 'image categorization': ['fine grain image categorization'],\n",
       " 'imitation learning': [],\n",
       " 'improved training for wasserstein gan': [],\n",
       " 'inductive learning': [],\n",
       " 'infogan': [],\n",
       " 'information theory': [],\n",
       " 'infrared and visible image fusion': [],\n",
       " 'instance search': [],\n",
       " 'instance segmentation': ['3d instance segmentation',\n",
       "  '3d semantic instance segmentation'],\n",
       " 'intelligent agent': [],\n",
       " 'intelligent surveillance': [],\n",
       " 'intent segmentation': [],\n",
       " 'interactive segmentation': [],\n",
       " 'interlingual mt': [],\n",
       " 'intermediate level vision': ['binary shape analysis',\n",
       "  'boundary pattern analysis',\n",
       "  'circle and ellipse detection',\n",
       "  'hough transform',\n",
       "  'line detection',\n",
       "  'pattern matching'],\n",
       " 'interpretable machine learning': [],\n",
       " 'intrinsic image decomposition': [],\n",
       " 'invariants and perspective': [],\n",
       " 'isolated gestures recognition': [],\n",
       " 'java': [],\n",
       " 'joke generation': [],\n",
       " 'jpeg compression artifact reduction': [],\n",
       " 'julia': [],\n",
       " 'k means clustering': [],\n",
       " 'knime': [],\n",
       " 'kxen': [],\n",
       " 'keras': [],\n",
       " 'kernel methods': ['svm'],\n",
       " 'kernel svm': [],\n",
       " 'keypoint localization': [],\n",
       " 'keyword spotting': [],\n",
       " 'knowledge engineering': [],\n",
       " 'knowledge graph embedding': [],\n",
       " 'knowledge and reasoning': ['logic based reasoning',\n",
       "  'default reasoning and the qualification problem',\n",
       "  'general ontology'],\n",
       " 'knowledge based': ['reinforcement learning'],\n",
       " 'known environment': [],\n",
       " 'lars lasso': [],\n",
       " 'lionsolver': [],\n",
       " 'lpboost': [],\n",
       " 'lstm': [],\n",
       " 'language based temporal localization': [],\n",
       " 'large scale person re identification': [],\n",
       " 'large vocabulary continuous speech recognition': [],\n",
       " 'lasso regression': [],\n",
       " 'latent semantic analysis': [],\n",
       " 'lenet': [],\n",
       " 'learning from observations': ['decision trees',\n",
       "  'general logical descriptions',\n",
       "  'inductive learning',\n",
       "  'information theory'],\n",
       " 'least angle regression': [],\n",
       " 'least squares gan': [],\n",
       " 'lemmatisation algorithms': [],\n",
       " 'lemmatization': [],\n",
       " 'lempel ziv welch algorithm': [],\n",
       " 'lexeme based morphology': [],\n",
       " 'lexical semantics': [],\n",
       " 'lexical recognition': [],\n",
       " 'liblinear': [],\n",
       " 'libsvm': [],\n",
       " 'lightgbm': [],\n",
       " 'light gbm': [],\n",
       " 'line detection': [],\n",
       " 'line segment detection': [],\n",
       " 'linear regression': ['elastic net',\n",
       "  'elastic net',\n",
       "  'ordinary least squares regression',\n",
       "  'stochastic gradient descent regression'],\n",
       " 'linguistics': ['analogy',\n",
       "  'hypothesis of creative patterns',\n",
       "  'joke generation',\n",
       "  'metaphor and simile',\n",
       "  'neologism',\n",
       "  'poetry',\n",
       "  'story generation'],\n",
       " 'lipreading': [],\n",
       " 'lisp': [],\n",
       " 'local color enhancement': [],\n",
       " 'local invariant feature detectors and descriptors': [],\n",
       " 'local outlier factor': [],\n",
       " 'logic based': [],\n",
       " 'logical reasoning': ['description logic',\n",
       "  'frame systems and semantic networks',\n",
       "  'indexing, retrieval and unification'],\n",
       " 'logistic regression': [],\n",
       " 'logitboost': [],\n",
       " 'loss function': ['classification loss functions',\n",
       "  'embedding loss functions',\n",
       "  'regression loss functions'],\n",
       " 'lossy compression artifact reduction': [],\n",
       " 'low level vision': ['basic image filtering',\n",
       "  'corner and interest point detection',\n",
       "  'edge detection',\n",
       "  'mathematical morphology',\n",
       "  'texture',\n",
       "  'thresholding'],\n",
       " 'low light image enhancement': [],\n",
       " 'low variance filter': [],\n",
       " 'magan': [],\n",
       " 'matlab': [],\n",
       " 'mlpack': [],\n",
       " 'moa': [],\n",
       " 'mx net': [],\n",
       " 'machine learning': ['data science',\n",
       "  'feature engineering',\n",
       "  'belief network',\n",
       "  'feature learning',\n",
       "  'knowledge based',\n",
       "  'learning from observations',\n",
       "  'neural networks',\n",
       "  'semi supervised learning',\n",
       "  'supervised learning',\n",
       "  'unsupervised learning'],\n",
       " 'machine translation': ['example based mt',\n",
       "  'hybrid mt',\n",
       "  'neural mt',\n",
       "  'rule based mt',\n",
       "  'statistical based mt'],\n",
       " 'machine vision': [],\n",
       " 'manner of articulation detection': [],\n",
       " 'markerless motion capture': [],\n",
       " 'market analysis and data mining': [],\n",
       " 'marketing forecasting': [],\n",
       " 'marketing and advertising': ['marketing',\n",
       "  'trade',\n",
       "  'ad targeting',\n",
       "  'content generation',\n",
       "  'dynamic pricing',\n",
       "  'fraudalent ads',\n",
       "  'marketing forecasting',\n",
       "  'predictive analytics for marketing',\n",
       "  'predictive customer service',\n",
       "  'products recognition',\n",
       "  'programmatic advertising',\n",
       "  'recommendation engine'],\n",
       " 'mask rcnn': [],\n",
       " 'matching algorithms': [],\n",
       " 'material classification': ['materials screening'],\n",
       " 'material recognition': [],\n",
       " 'materials screening': [],\n",
       " 'mathematica': [],\n",
       " 'mathematical morphology': [],\n",
       " 'mathematical question answering': [],\n",
       " 'mean shift clustering': [],\n",
       " 'medication management': [],\n",
       " 'meta learning': [],\n",
       " 'metamerism': [],\n",
       " 'metaphor and simile': [],\n",
       " 'methods based on projections': [],\n",
       " 'metric learning': [],\n",
       " 'microsoft cntk': [],\n",
       " 'military': [],\n",
       " 'missing markers reconstruction': [],\n",
       " 'missing value ratio': [],\n",
       " 'mixture model': [],\n",
       " 'modal logic': [],\n",
       " 'mode regularized gan': [],\n",
       " 'model compression': [],\n",
       " 'model selection': [],\n",
       " 'monocular depth estimation': [],\n",
       " 'montecarlomachinelearninglibrary': [],\n",
       " 'morpheme based morphology': [],\n",
       " 'morphological segmentation': ['lexeme based morphology',\n",
       "  'morpheme based morphology',\n",
       "  'word based morphology'],\n",
       " 'motion': ['optical flow',\n",
       "  'stereo from motion',\n",
       "  'the kalman filter',\n",
       "  'time to adjacency analysis',\n",
       "  'wide baseline matching'],\n",
       " 'motion capture': ['markerless motion capture',\n",
       "  'missing markers reconstruction'],\n",
       " 'motion estimation': [],\n",
       " 'motion planning': [],\n",
       " 'motion segmentation': [],\n",
       " 'multi goal reinforcement learning': [],\n",
       " 'multi label image retrieval': [],\n",
       " 'multi layer perceptron': [],\n",
       " 'multi oriented scene text detection': [],\n",
       " 'multi task learning': [],\n",
       " 'multi agent reinforcement learning': [],\n",
       " 'multi modal image segmentation': [],\n",
       " 'multilayer feed forward networks': ['cnn',\n",
       "  'multi layer perceptron',\n",
       "  'radial basis network'],\n",
       " 'multimodal emotion recognition': [],\n",
       " 'multimodal unsupervised image to image translation': [],\n",
       " 'multiple action detection': [],\n",
       " 'multiple instance learning': [],\n",
       " 'multiple object tracking': [],\n",
       " 'multiview learning': [],\n",
       " 'music genre transfer': [],\n",
       " 'musical': ['music'],\n",
       " 'naive bayes classifier': [],\n",
       " 'named entity recognition': ['conditional random fields approach'],\n",
       " 'natural language generation': [],\n",
       " 'natural language processing': ['text mining',\n",
       "  'atc',\n",
       "  'discourse',\n",
       "  'semantics',\n",
       "  'speech',\n",
       "  'syntax'],\n",
       " 'natural language understanding': [],\n",
       " 'nearest neighbor': [],\n",
       " 'neologism': [],\n",
       " 'network embedding': [],\n",
       " 'network pruning': [],\n",
       " 'neural network based': ['deep learning'],\n",
       " 'neural architecture search': [],\n",
       " 'neural mt': [],\n",
       " 'neural networks': ['artificial neural networks'],\n",
       " 'neurosolutions': [],\n",
       " 'no reference image quality assessment': [],\n",
       " 'noisy speech recognition': [],\n",
       " 'non monotonic logic': [],\n",
       " 'non optimization based': ['k nn'],\n",
       " 'novel view synthesis': [],\n",
       " 'ocr': [],\n",
       " 'optics algorithm': [],\n",
       " 'object classification': [],\n",
       " 'object counting': [],\n",
       " 'object detection': ['salient object detection',\n",
       "  'weakly supervised object detection'],\n",
       " 'object labeling and counting': [],\n",
       " 'object localization': ['active object localization',\n",
       "  'image based localization',\n",
       "  'unsupervised object localization',\n",
       "  'weakly supervised object localization'],\n",
       " 'object recognition': ['continuous object recognition',\n",
       "  'depiction invariant object recognition',\n",
       "  '3d object recognition'],\n",
       " 'object reconstruction': [],\n",
       " 'object tracking': ['multiple object tracking',\n",
       "  'thermal infrared object tracking',\n",
       "  'visual object tracking'],\n",
       " 'observation completion': [],\n",
       " 'octave': [],\n",
       " 'one class svm': [],\n",
       " 'one shot learning': [],\n",
       " 'ontology': [],\n",
       " 'opencv': [],\n",
       " 'opennn': [],\n",
       " 'open domain question answering': [],\n",
       " 'optical character recognition': ['handwritten digit recognition'],\n",
       " 'optical flow': [],\n",
       " 'optical flow estimation': [],\n",
       " 'oracledatamining': [],\n",
       " 'orange': [],\n",
       " 'ordinary least squares regression': [],\n",
       " 'orthogonal matching pursuit': [],\n",
       " 'out of distribution detection': [],\n",
       " 'outlier detection': [],\n",
       " 'pca': [],\n",
       " 'parsing': [],\n",
       " 'partial order planning': [],\n",
       " 'partial domain adaptation': [],\n",
       " 'partially instantiated operators': [],\n",
       " 'passive aggressive algorithms': [],\n",
       " 'passive learning': ['known environment', 'unknown environment'],\n",
       " 'pattern matching': [],\n",
       " 'pattern recognition': ['kernel methods', 'statistical methods'],\n",
       " 'pattern theory': [],\n",
       " 'pedestrian detection': [],\n",
       " 'perception': ['computer vision', 'speech'],\n",
       " 'perceptrons': [],\n",
       " 'perl': [],\n",
       " 'person re identification': ['large scale person re identification',\n",
       "  'semi supervised person re identification',\n",
       "  'unsupervised person re identification',\n",
       "  'video based person re identification'],\n",
       " 'person retrieval': [],\n",
       " 'personal finance': [],\n",
       " 'phonotactic cues': [],\n",
       " 'photo to caricature translation': [],\n",
       " 'planning': ['conditional planning',\n",
       "  'expressive operator descriptions',\n",
       "  'hierarchical decomposition',\n",
       "  'intelligent agent',\n",
       "  'knowledge engineering',\n",
       "  'partial order planning',\n",
       "  'partially instantiated operators'],\n",
       " 'pos tagging': ['brill tagger', 'structure regularization method'],\n",
       " 'poetry': [],\n",
       " 'policy gradient methods': [],\n",
       " 'portfolio management': [],\n",
       " 'pose estimation': [],\n",
       " 'predictive analytics for marketing': [],\n",
       " 'predictive domain adaptation': [],\n",
       " 'predictive customer service': [],\n",
       " 'probabilistic programming': [],\n",
       " 'probabilistic methods for uncertain reasoning': [],\n",
       " 'problem solving': [],\n",
       " 'products recognition': [],\n",
       " 'programmatic advertising': [],\n",
       " 'projection pursuit': [],\n",
       " 'prolog': [],\n",
       " 'proportional logic': [],\n",
       " 'provide consultations and companion robots': [],\n",
       " 'python': [],\n",
       " 'pytorch': [],\n",
       " 'q learning': [],\n",
       " 'quantization': [],\n",
       " 'question answering': ['mathematical question answering',\n",
       "  'open domain question answering'],\n",
       " 'rbm': ['binary rbm with contrastive divergence',\n",
       "  'binary rbm with persistent contrastive divergence'],\n",
       " 'rcase': [],\n",
       " 'rcnn': [],\n",
       " 'rnn': ['bidirectional rnn', 'deep rnn', 'gated recurrent unit', 'lstm'],\n",
       " 'r language': [],\n",
       " 'radial basis network': [],\n",
       " 'radial basis svm': [],\n",
       " 'random forest': [],\n",
       " 'random indexing': [],\n",
       " 'rapidminer': [],\n",
       " 'recommendation engine': [],\n",
       " 'recurrent neural networks': [],\n",
       " 'recursive neural networks': [],\n",
       " 'regression': ['bayesian regression',\n",
       "  'decision tree regressor',\n",
       "  'gradient boost regressor',\n",
       "  'lars lasso',\n",
       "  'lasso regression',\n",
       "  'least angle regression',\n",
       "  'linear regression',\n",
       "  'orthogonal matching pursuit',\n",
       "  'passive aggressive algorithms',\n",
       "  'ridge regression',\n",
       "  'robustness regression'],\n",
       " 'reinforcement learning': ['active learning',\n",
       "  'deep adversarial networks',\n",
       "  'passive learning',\n",
       "  'q learning',\n",
       "  'temporal difference'],\n",
       " 'relationship extraction': [],\n",
       " 'representation learning': ['document embedding',\n",
       "  'graph embedding',\n",
       "  'graph representation learning',\n",
       "  'knowledge graph embedding',\n",
       "  'network embedding',\n",
       "  'sentence embeddings',\n",
       "  'unsupervised representation learning',\n",
       "  'word embeddings'],\n",
       " 'resnet': [],\n",
       " 'ridge regression': [],\n",
       " 'robot navigation': [],\n",
       " 'robot task planning': [],\n",
       " 'robotic grasping': [],\n",
       " 'robotics': ['deformable object manipulation',\n",
       "  'effector',\n",
       "  'motion planning',\n",
       "  'robot navigation',\n",
       "  'robot task planning',\n",
       "  'robotic grasping',\n",
       "  'sensor',\n",
       "  'visual navigation',\n",
       "  'visual odometry',\n",
       "  'language based assistance'],\n",
       " 'robust speech recognition': [],\n",
       " 'robustness regression': ['huber regression',\n",
       "  'theil sen estimator',\n",
       "  'random sample consensus(ransac)'],\n",
       " 'ruby': [],\n",
       " 'rule based mt': ['dictionary based mt',\n",
       "  'interlingual mt',\n",
       "  'transfer based mt'],\n",
       " 'sap': [],\n",
       " 'sas': [],\n",
       " 'sbd': [],\n",
       " 'spss': [],\n",
       " 'sqlserveranalysisservices': [],\n",
       " 'statistica': [],\n",
       " 'svd': [],\n",
       " 'svm': [],\n",
       " 'saliency detection': ['co saliency detection', 'video saliency detection'],\n",
       " 'saliency prediction': [],\n",
       " 'salient object detection': [],\n",
       " 'salt and pepper noise removal': [],\n",
       " 'sammon mapping': [],\n",
       " 'scene aware dialogue': [],\n",
       " 'scene classification': [],\n",
       " 'scene flow estimation': [],\n",
       " 'scene graph generation': [],\n",
       " 'scene parsing': ['scene graph generation',\n",
       "  'scene recognition',\n",
       "  'scene text recognition',\n",
       "  'scene understanding'],\n",
       " 'scene recognition': [],\n",
       " 'scene segmentation': [],\n",
       " 'scene text detection': ['curved text detection',\n",
       "  'multi oriented scene text detection'],\n",
       " 'scene text recognition': [],\n",
       " 'scene understanding': [],\n",
       " 'scruffy': [],\n",
       " 'search and optimization': ['optimization algorithms', 'search algorithms'],\n",
       " 'self organizing maps': [],\n",
       " 'semantic segmentation': ['weakly supervised semantic segmentation'],\n",
       " 'semantic folding': [],\n",
       " 'semantics': ['distributional semantics',\n",
       "  'lexical semantics',\n",
       "  'machine translation',\n",
       "  'named entity recognition',\n",
       "  'natural language generation',\n",
       "  'natural language understanding',\n",
       "  'ocr',\n",
       "  'question answering',\n",
       "  'relationship extraction',\n",
       "  'sentiment analysis',\n",
       "  'textual entailment recognition',\n",
       "  'topic segmentation',\n",
       "  'word sense disambiguation'],\n",
       " 'semi supervised image classification': [],\n",
       " 'semi supervised learning': [],\n",
       " 'semi supervised person re identification': [],\n",
       " 'semi supervised video object segmentation': [],\n",
       " 'sensor': [],\n",
       " 'sensor modeling': [],\n",
       " 'sentence embeddings': [],\n",
       " 'sentence segmentaion': [],\n",
       " 'sentiment analysis': [],\n",
       " 'sequence to sequence speecg recognition': [],\n",
       " 'sequential learning': [],\n",
       " 'sequitur': [],\n",
       " 'session based recommendations': [],\n",
       " 'shape recognition': [],\n",
       " 'shogun': [],\n",
       " 'sign language recognition': [],\n",
       " 'sign language translation': [],\n",
       " 'simultaneous localization and mapping': [],\n",
       " 'sinewave synthesis': [],\n",
       " 'single image blind deblurring': [],\n",
       " 'size filtering': [],\n",
       " 'skeleton based action recognition': [],\n",
       " 'skeletons and thinning': [],\n",
       " 'sketch based image retrieval': [],\n",
       " 'sketch recognition': [],\n",
       " 'social intelligence': [],\n",
       " 'soft computing': ['computational creativity',\n",
       "  'evolutionary computation',\n",
       "  'fuzzy systems'],\n",
       " 'softmax gan': [],\n",
       " 'spark mlib': [],\n",
       " 'sparse learning': [],\n",
       " 'spatio temporal action localization': [],\n",
       " 'speaker identification': [],\n",
       " 'speaker recognition': [],\n",
       " 'speaking style synthesis': [],\n",
       " 'spectral clustering': [],\n",
       " 'spectral estimation': ['spectrum cartography'],\n",
       " 'spectrum cartography': [],\n",
       " 'speech': ['speech processing',\n",
       "  'acoustic modelling',\n",
       "  'acoustic question answering',\n",
       "  'audio visual speech recognition',\n",
       "  'denoising',\n",
       "  'emotion recognition',\n",
       "  'keyword spotting',\n",
       "  'large vocabulary continuous speech recognition',\n",
       "  'manner of articulation detection',\n",
       "  'speaker identification',\n",
       "  'speaker recognition',\n",
       "  'speaking style synthesis',\n",
       "  'speech recognition',\n",
       "  'speech segmentation',\n",
       "  'speech enhancement',\n",
       "  'speech recognition',\n",
       "  'speech separation',\n",
       "  'speech synthesis',\n",
       "  'speech verification',\n",
       "  'spoken dialogue systems',\n",
       "  'spoken language understanding',\n",
       "  'text independent speaker recognition',\n",
       "  'text to speech synthesis',\n",
       "  'text to speech',\n",
       "  'voice conversion',\n",
       "  'voice query recognition'],\n",
       " 'speech recognition': ['accented speech recognition',\n",
       "  'conversational speech recognition',\n",
       "  'distant speech recognition',\n",
       "  'noisy speech recognition',\n",
       "  'robust speech recognition',\n",
       "  'sequence to sequence speecg recognition',\n",
       "  'visual speech recognition'],\n",
       " 'speech segmentation': ['lexical recognition', 'phonotactic cues'],\n",
       " 'speech emotion recognition': [],\n",
       " 'speech enhancement': [],\n",
       " 'speech separation': [],\n",
       " 'speech synthesis': [],\n",
       " 'speech verification': [],\n",
       " 'spoken dialogue systems': [],\n",
       " 'spoken language understanding': [],\n",
       " 'stacking': [],\n",
       " 'stata': [],\n",
       " 'statistical methods': ['classification methods',\n",
       "  'clustering methods',\n",
       "  'dimensionality reduction methods'],\n",
       " 'statistical based mt': [],\n",
       " 'statistical learning': [],\n",
       " 'steering control': [],\n",
       " 'stemming': ['affix stemmers',\n",
       "  'lemmatisation algorithms',\n",
       "  'matching algorithms',\n",
       "  'stochastic algorithms',\n",
       "  'suffix stripping algorithms',\n",
       "  'n gram analysis'],\n",
       " 'stereo depth estimation': [],\n",
       " 'stereo matching': [],\n",
       " 'stereo from motion': [],\n",
       " 'stochastic gradient descent regression': [],\n",
       " 'stochastic optimization': [],\n",
       " 'stochastic algorithms': [],\n",
       " 'story generation': [],\n",
       " 'style generalization': [],\n",
       " 'style transfer': ['image stylization'],\n",
       " 'sub symbolic based ai': ['behavior based ai',\n",
       "  'nouvelle ai',\n",
       "  'statistical ai',\n",
       "  'computational intelligence',\n",
       "  'embodied intelligence',\n",
       "  'soft computing'],\n",
       " 'suffix stripping algorithms': [],\n",
       " 'super resolution': ['depth map super resolution',\n",
       "  'image super resolution',\n",
       "  'video super resolution',\n",
       "  '3d object super resolution'],\n",
       " 'supervised learning': ['classification',\n",
       "  'ensemble learning',\n",
       "  'non optimization based',\n",
       "  'regression'],\n",
       " 'supervised video summarization': [],\n",
       " 'support vector machine': ['kernel svm', 'radial basis svm'],\n",
       " 'swarm intelligence': [],\n",
       " 'symbolic based ai': ['cognitive simulation',\n",
       "  'knowledge based',\n",
       "  'logic based',\n",
       "  'scruffy'],\n",
       " 'syntax': ['grammar induction',\n",
       "  'lemmatization',\n",
       "  'morphological segmentation',\n",
       "  'parsing',\n",
       "  'pos tagging',\n",
       "  'sbd',\n",
       "  'stemming',\n",
       "  'terminology extraction',\n",
       "  'text segmentaion'],\n",
       " 'synthetic to real translation': [],\n",
       " 'tasks': ['action classification',\n",
       "  'action detection',\n",
       "  'action localization',\n",
       "  'activity recognition',\n",
       "  'animation',\n",
       "  'anomaly detection',\n",
       "  'art analysis',\n",
       "  'audio visual synchronization',\n",
       "  'autonomous vehicles',\n",
       "  'boundary detection',\n",
       "  'color constancy',\n",
       "  'colorization',\n",
       "  'conditional image generation',\n",
       "  'contour detection',\n",
       "  'deblurring',\n",
       "  'deep attention',\n",
       "  'defocus estimation',\n",
       "  'demosaicking',\n",
       "  'denoising',\n",
       "  'dense pixel correspondence estimation',\n",
       "  'depth completion',\n",
       "  'depth estimation',\n",
       "  'disparity estimation',\n",
       "  'document image classification',\n",
       "  'document to image conversion',\n",
       "  'drawing pictures',\n",
       "  'dynamic texture recognition',\n",
       "  'edge detection',\n",
       "  'event based vision',\n",
       "  'face transfer',\n",
       "  'facial recognition and modelling',\n",
       "  'fake image detection',\n",
       "  'few shot learning',\n",
       "  'forgery',\n",
       "  'gaze estimation',\n",
       "  'gesture recognition',\n",
       "  'homography estimation',\n",
       "  'human object interaction detection',\n",
       "  'hyperspectral',\n",
       "  'image to image translation',\n",
       "  'image captioning',\n",
       "  'image classification',\n",
       "  'image clustering',\n",
       "  'image comprehension',\n",
       "  'image compression',\n",
       "  'image cropping',\n",
       "  'image declipping',\n",
       "  'image deconvolution',\n",
       "  'image dehazing',\n",
       "  'image enhancement',\n",
       "  'image generation',\n",
       "  'image imputation',\n",
       "  'image inpainting',\n",
       "  'image manipulation detection',\n",
       "  'image matching',\n",
       "  'image matting',\n",
       "  'image quality assessment',\n",
       "  'image recognition',\n",
       "  'image reconstruction',\n",
       "  'image registration',\n",
       "  'image restoration',\n",
       "  'image retrieval',\n",
       "  'image similarity search',\n",
       "  'image steganography',\n",
       "  'image stitching',\n",
       "  'image categorization',\n",
       "  'infrared and visible image fusion',\n",
       "  'instance search',\n",
       "  'instance segmentation',\n",
       "  'intelligent surveillance',\n",
       "  'interactive segmentation',\n",
       "  'intrinsic image decomposition',\n",
       "  'line segment detection',\n",
       "  'lipreading',\n",
       "  'material classification',\n",
       "  'material recognition',\n",
       "  'metamerism',\n",
       "  'motion capture',\n",
       "  'motion estimation',\n",
       "  'motion segmentation',\n",
       "  'multi modal image segmentation',\n",
       "  'multiview learning',\n",
       "  'music genre transfer',\n",
       "  'novel view synthesis',\n",
       "  'object classification',\n",
       "  'object counting',\n",
       "  'object detection',\n",
       "  'object localization',\n",
       "  'object recognition',\n",
       "  'object reconstruction',\n",
       "  'object tracking',\n",
       "  'observation completion',\n",
       "  'one shot learning',\n",
       "  'optical flow estimation',\n",
       "  'out of distribution detection',\n",
       "  'person re identification',\n",
       "  'person retrieval',\n",
       "  'photo to caricature translation',\n",
       "  'pose estimation',\n",
       "  'saliency detection',\n",
       "  'saliency prediction',\n",
       "  'scene aware dialogue',\n",
       "  'scene classification',\n",
       "  'scene flow estimation',\n",
       "  'scene parsing',\n",
       "  'scene segmentation',\n",
       "  'scene text detection',\n",
       "  'semantic segmentation',\n",
       "  'sensor modeling',\n",
       "  'sign language recognition',\n",
       "  'sign language translation',\n",
       "  'simultaneous localization and mapping',\n",
       "  'sketch recognition',\n",
       "  'spectral estimation',\n",
       "  'stereo matching',\n",
       "  'style generalization',\n",
       "  'style transfer',\n",
       "  'super resolution',\n",
       "  'temporal localization',\n",
       "  'text to image generation',\n",
       "  'text spotting',\n",
       "  'texture classification',\n",
       "  'texture synthesis',\n",
       "  'trajectory prediction',\n",
       "  'unbalanced segmentation',\n",
       "  'vnla',\n",
       "  'video captioning',\n",
       "  'video classification',\n",
       "  'video inpainting',\n",
       "  'video object segmentation',\n",
       "  'video prediction',\n",
       "  'video quality assessment',\n",
       "  'video question answering',\n",
       "  'video summarization',\n",
       "  'viewpoint estimation',\n",
       "  'vision language navigation',\n",
       "  'visual crowd analysis',\n",
       "  'visual localization',\n",
       "  'visual question answering',\n",
       "  'visual sentiment prediction',\n",
       "  'zero shot action recognition',\n",
       "  'zero shot learning',\n",
       "  'image outpainting',\n",
       "  '3d',\n",
       "  'image/document clustering',\n",
       "  'keypoint detection/interest point detection',\n",
       "  'optical character recognition (ocr)'],\n",
       " 'temporal action proposal generation': [],\n",
       " 'temporal difference': [],\n",
       " 'temporal localization': ['camera localization',\n",
       "  'camera relocalization',\n",
       "  'language based temporal localization'],\n",
       " 'tensorflow': [],\n",
       " 'terminology extraction': [],\n",
       " 'text independent speaker recognition': [],\n",
       " 'text to speech synthesis': [],\n",
       " 'text to image generation': [],\n",
       " 'text to speech': ['articulatory synthesis',\n",
       "  'concatenation synthesis',\n",
       "  'deep learning synthesis',\n",
       "  'formant synthesis',\n",
       "  'hmm based synthesis',\n",
       "  'sinewave synthesis'],\n",
       " 'text segmentaion': ['intent segmentation',\n",
       "  'sentence segmentaion',\n",
       "  'topic segmentation',\n",
       "  'word segmentation'],\n",
       " 'text spotting': [],\n",
       " 'textual entailment recognition': [],\n",
       " 'texture': [],\n",
       " 'texture classification': [],\n",
       " 'texture synthesis': [],\n",
       " 'the kalman filter': [],\n",
       " 'theano': [],\n",
       " 'theil sen estimator': [],\n",
       " 'thematics': ['natural language processing', 'perception', 'robotics'],\n",
       " 'thermal infrared object tracking': [],\n",
       " 'thresholding': ['entropy based thresholding',\n",
       "  'global valley approach',\n",
       "  'variance based thresholding',\n",
       "  'adaptive thresholding'],\n",
       " 'time to adjacency analysis': [],\n",
       " 'topic segmentation': [],\n",
       " 'trajectory prediction': [],\n",
       " 'transfer based mt': [],\n",
       " 'transfer reinforcement learning': [],\n",
       " 'tribox method': [],\n",
       " 'umap': [],\n",
       " 'unbalanced segmentation': [],\n",
       " 'uncertain knowledge and reasoning': ['dempster shafer theory',\n",
       "  'fuzzy sets',\n",
       "  'bayesian network',\n",
       "  'decision theory',\n",
       "  'dynamic belief networks',\n",
       "  'dynamic decision networks',\n",
       "  'hidden markov models',\n",
       "  'probabilistic methods for uncertain reasoning',\n",
       "  'utility theory'],\n",
       " 'underwriting': [],\n",
       " 'unit selection synthesis': [],\n",
       " 'unknown environment': [],\n",
       " 'unsupervised anomaly detection': [],\n",
       " 'unsupervised anomaly detection in sound': [],\n",
       " 'unsupervised domain adaptation': [],\n",
       " 'unsupervised image to image translation': ['multimodal unsupervised image to image translation',\n",
       "  'photo to caricature translation',\n",
       "  'synthetic to real translation'],\n",
       " 'unsupervised learning': ['anamoly detection',\n",
       "  'clustering',\n",
       "  'dimensionality reduction'],\n",
       " 'unsupervised object localization': [],\n",
       " 'unsupervised person re identification': [],\n",
       " 'unsupervised representation learning': [],\n",
       " 'unsupervised video object segmentation': [],\n",
       " 'unsupervised video summarization': [],\n",
       " 'utility theory': [],\n",
       " 'vggnet': [],\n",
       " 'vnla': [],\n",
       " 'vanilla gan': [],\n",
       " 'vanilla vae': [],\n",
       " 'variational autoencoder': ['adversarial autoencoder',\n",
       "  'adversarial variational bayes',\n",
       "  'conditional vae',\n",
       "  'denoising vae',\n",
       "  'vanilla vae'],\n",
       " 'video based person re identification': [],\n",
       " 'video captioning': ['audio visual video captioning',\n",
       "  'dense video captioning'],\n",
       " 'video classification': [],\n",
       " 'video inpainting': [],\n",
       " 'video object segmentation': ['semi supervised video object segmentation',\n",
       "  'unsupervised video object segmentation'],\n",
       " 'video prediction': [],\n",
       " 'video quality assessment': [],\n",
       " 'video question answering': [],\n",
       " 'video saliency detection': [],\n",
       " ...}"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "adjList"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('brute force search', 0.19183763624449512),\n",
       " ('depth first search', 0.2423863112190886),\n",
       " ('uniform cost search', 0.24792623709578712),\n",
       " ('breadth first search', 0.28157797099336207),\n",
       " ('bidirectional search', 0.3022612707588843),\n",
       " ('iterative deepening depth first search', 0.37466425974563805),\n",
       " ('blind search', 0.7110479923891726),\n",
       " ('tabu search algorithm', 0.734821229458574),\n",
       " ('search algorithms', 0.7509061587479332),\n",
       " ('gravitational search algorithm', 0.7789498463447094)]"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "[('brute force search', 0.19183763624449512),\n",
       " ('depth first search', 0.2423863112190886),\n",
       " ('uniform cost search', 0.24792623709578712),\n",
       " ('breadth first search', 0.28157797099336207),\n",
       " ('bidirectional search', 0.3022612707588843),\n",
       " ('iterative deepening depth first search', 0.37466425974563805),\n",
       " ('blind search', 0.7110479923891726),\n",
       " ('tabu search algorithm', 0.734821229458574),\n",
       " ('search algorithms', 0.7509061587479332),\n",
       " ('gravitational search algorithm', 0.7789498463447094)]"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.kv.most_similar((model.kv.get_vector('breadth first search') + model.kv.get_vector('depth first search')) / 2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('adam', 0.014114604754161875),\n",
       " ('adagrad', 0.014183437772351783),\n",
       " ('adadelta', 0.029745172911442),\n",
       " ('first order optimization algorithms', 0.039091551339758855),\n",
       " ('mini batch gradient descent', 0.663839885552896),\n",
       " ('nesterov accelerated gradient', 0.6787875250172928),\n",
       " ('gradient descent', 0.6886457532065207),\n",
       " ('stochastic gradient descent', 0.7253832900773263),\n",
       " ('second order optimization algorithms', 1.0734767399911862),\n",
       " ('optimization algorithms', 1.084217470829965)]"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "[('adam', 0.014114604754161875),\n",
       " ('adagrad', 0.014183437772351783),\n",
       " ('adadelta', 0.029745172911442),\n",
       " ('first order optimization algorithms', 0.039091551339758855),\n",
       " ('mini batch gradient descent', 0.663839885552896),\n",
       " ('nesterov accelerated gradient', 0.6787875250172928),\n",
       " ('gradient descent', 0.6886457532065207),\n",
       " ('stochastic gradient descent', 0.7253832900773263),\n",
       " ('second order optimization algorithms', 1.0734767399911862),\n",
       " ('optimization algorithms', 1.084217470829965)]"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.kv.most_similar((model.kv.get_vector('adam') + model.kv.get_vector('adagrad')) / 2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('binary sigmoidal function', 0.18241587641269197),\n",
       " ('bipolar sigmoidal function', 0.1863036112783289),\n",
       " ('sigmoidal function', 0.30586660940945043),\n",
       " ('softstep function', 0.3181918899998072),\n",
       " ('logistic function', 0.3612456003548927),\n",
       " ('maxout function', 0.5592107072006494),\n",
       " ('ramp function', 0.5592362361320333),\n",
       " ('bipolar step function', 0.5596667536528842),\n",
       " ('leaky rectified linear unit (leaky relu)', 0.5611363902813842),\n",
       " ('sigmoid linear unit (silu)', 0.5618453955212401)]"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "[('binary sigmoidal function', 0.18241587641269197),\n",
       " ('bipolar sigmoidal function', 0.1863036112783289),\n",
       " ('sigmoidal function', 0.30586660940945043),\n",
       " ('softstep function', 0.3181918899998072),\n",
       " ('logistic function', 0.3612456003548927),\n",
       " ('maxout function', 0.5592107072006494),\n",
       " ('ramp function', 0.5592362361320333),\n",
       " ('bipolar step function', 0.5596667536528842),\n",
       " ('leaky rectified linear unit (leaky relu)', 0.5611363902813842),\n",
       " ('sigmoid linear unit (silu)', 0.5618453955212401)]"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.kv.most_similar('sigmoid function')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "vec = embed(['hello'])[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1024,)"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "vec.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[1, 2, 3, 4, 5, 6]"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "[1, 2, 3]+ [4, 5, 6]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}