SPPU · Final Year Computer Engineering · 2019 Course (Revised)
Natural Language Processing
410252A · Elective V · Semester VIII · 3 credits · In-semester 30 + end-semester 70
One of the Elective V options.
Unit 1: Introduction to Natural Language Processing
7 hoursWhy NLP is hard, programming vs natural languages, finite automata for NLP, stages of NLP, open problems; tokenisation, stemming, lemmatisation, part-of-speech tagging.
Marks: 15 in the in-semester exam
- Natural language processing and its challenges
- Stages of NLP
- Tokenisation, stemming and lemmatisation
- Part-of-speech tagging
Unit 2: Language Syntax and Semantics
7 hoursMorphology, inflectional and derivational, parsing with finite-state transducers; syntactic representations, parsing algorithms, PCFGs and statistical parsing; lexical semantics, homonymy, polysemy, synonymy, hyponymy, WordNet, WSD, latent semantic analysis.
Marks: 15 in the in-semester exam
- Morphological analysis and finite-state transducers
- Syntactic parsing and probabilistic CFGs
- Lexical semantics and WordNet
- Word sense disambiguation
Unit 3: Language Modelling
7 hoursProbabilistic, Markov, generative, log-linear and graph-based models; n-grams, smoothing, evaluation; bag-of-words, TF-IDF, word2vec, doc2vec, BERT; LDA, LSA and NMF topic models.
Marks: 17–18 in the end-semester exam
- N-gram language models and smoothing
- Evaluating language models
- Bag-of-words and TF-IDF
- Word embeddings: word2vec and doc2vec
- Contextual embeddings and BERT
- Topic modelling: LDA, LSA and NMF
Unit 4: Information Retrieval using NLP
7 hoursVector space retrieval; named entity recognition and its evaluation; entity and relation extraction; reference and coreference resolution; cross-lingual IR.
Marks: 17–18 in the end-semester exam
- Information retrieval with the vector space model
- Named entity recognition
- Relation extraction
- Coreference resolution
- Cross-lingual information retrieval
Unit 5: NLP Tools and Techniques
8 hoursNLTK, spaCy, TextBlob and Gensim; WordNets, IndoWordNet, VerbNet, PropBank, treebanks and universal dependencies; the Lesk and Walker algorithms for WSD.
Marks: 17–18 in the end-semester exam
- NLP libraries: NLTK, spaCy, TextBlob and Gensim
- Lexical resources: WordNet, IndoWordNet, VerbNet and PropBank
- Treebanks and universal dependencies
- Lesk and Walker's algorithms
Unit 6: Applications of NLP
7 hoursRule-based and statistical machine translation, cross-lingual translation; sentiment analysis, question answering, textual entailment, discourse processing, dialogue agents, natural language generation.
Marks: 17–18 in the end-semester exam
- Machine translation
- Sentiment analysis
- Question answering and textual entailment
- Dialogue systems and natural language generation
Unit contents are summarised from the official SPPU syllabus . AiCanCode.org is not affiliated with Savitribai Phule Pune University.