Final Exam: Natural Language Processing

WHAT YOU WILL LEARN

• 

  Illustrate phonomes, morpheme and lexemes
  

  describe syntactic and semantic analysis for nlp
  

  describe heuristic approaches to solve nlp tasks
  

  specify challenges with nlp in real world problem solving
  

  illustrate various tools in nlp used across different industries
  

  explain basic overview of nltk ecosystem
  

  classify the difference between spacy and nltk
  

  explain overview of spacy models and language support
  

  perform spacy installations and models, part of speech, morphology, lemmatization; demonstrate dependency parsing, named entities, entity linking, tokenization, merging & splitting; demonstrate sentence segmentation, sentence similarity,
  

  illustrate and extract synonyms and identify wordnet hierarchies - hypernyms and hyponyms
  

  demonstrate python re module, re - search, find all, finditer, groups, find and replace, split
  

  recall anchors, character classes, greedy, lazy and backtracking, performance
  

  identify anchors, character classes, greedy, lazy and backtracking, performance
  

  perform exploration of meronyms and holonymms
  

  illustrate one-hot-encoding, bag of words, n-gram, tfidf
  

  demonstrate data loading and basic overview of columns
  

  demonstrate simple model building & evaluation
  

  restate logistic regression, svm, naive bayes & boosting models
  

  explain the existence of polyglot & textblob and describe the benefits over nltk & spacy, use-cases
  

  explain the existence of gensim
  

  demonstrate query similarity using gensim
  

  demonstrate building an lda model for topic modeling using genism
  

  explain nlp with deep learning;
  

  illustrate various use cases in nlp across different industries
  

  explain the basic overview of spacy, tensorflow ecosystem
  

  illustrate overview of data
  

  illustrate the basics of data loading and columns
  

  illustrate single layer perceptron architecture of neural network; illustrate multi layer perceptron architecture of neural network
  

  describe recurrent neural network architecture and how it can capture context in language
  

  illustrate different application of basic neural network based architecture
• 

  describe the steps to load the amazon product reviews dataset into google colaboratory
  

  explain the importance of memory based learning and different networks it supports
  

  perform review classification using gru
  

  describe gated recurrent unit and it's difference from rnn
  

  illustrate different types of lstm networks
  

  explain what is transfer learning and how it help's to get better results, explore different types of learnings and different types of transfer learning; specify advantages & challenges of transfer learning in real world problem solving
  

  describe and explore fasttext & word2vec
  

  implement and fine tune the lm model using ulmfit
  

  describe and explore elmo
  

  describe and explore fasttext & word2vec
  

  explain github bug data and problem statement; perform library loading, data loading and basic overview of columns
  

  demonstrate exploratory data analysis (eda) - word count analysis and label analysis
  

  explain github bug data and problem statement
  

  demonstrate eda - punctuation analysis, stop word analysis, and word cloud
  

  demonstrate fundamentals of transformer architecture
  

  discover the use of the attention method to improve nlp model results
  

  explain the encoder-decoder network architecture and its application
  

  demonstrate fundamentals of the self-attention layer in transformer architecture
  

  explore bert transfer learning approaches
  

  perform model evaluation and predicting sentiment for raw text
  

  discover different types of bert models and their significance
  

  perform data pre-processing and bert tokenization
  

  discover fundamentals of language models, explain basics of gpt models
  

  describe few shot learning concept used in gpt 3
  

  discover different industry use cases and challenges for gpt
  

  greedy search, beam search, and basic sampling
  

  explore machine translation problem and its application
  

  transformer encoder to produce a new representation of input sequence
  

  frame the problem end to end. understand how single sentence translation works
  

  transformer decoder to produce output target sequence