SPPU · Third Year Artificial Intelligence and Data Science · 2024 Pattern
Natural Language Processing
PCC-303-AID · Semester V · 3 credits · CCE 30 + end-semester 70
Unit 1: Linguistic and Formal Foundations
9 hoursHistory, levels, stages and applications of NLP; ambiguity; regular expressions and finite automata; context-free grammars, derivations and parse trees; morphology and finite-state transducers; minimum edit distance and noisy-channel spelling correction; rule-, data- and knowledge-based approaches.
Marks: 14 in the end-semester exam · CCE: Unit test (12 marks, Units 1 & 2)
- Levels and challenges of NLP
- Regular expressions and finite automata
- Context-free grammars
- Morphology and finite-state transducers
- Minimum edit distance
Unit 2: Text Representation
9 hoursCorpora and annotation; tokenisation, segmentation and normalisation; stemming (Porter) and lemmatisation; stop words; out-of-vocabulary words and subwords; one-hot and distributional representations; sparse versus dense vectors; WordNet; n-gram features.
Marks: 14 in the end-semester exam · CCE: Unit test (12 marks, Units 1 & 2)
- Corpora and annotation
- Tokenisation and normalisation
- Stemming and lemmatisation
- Subword tokenisation
- Word representations
- WordNet
Unit 3: Statistical NLP and Language Models
9 hoursConditional probability and maximum likelihood; unigram, bigram and trigram models, perplexity, backoff and interpolation; naive Bayes and logistic regression for text; HMMs and the Viterbi algorithm; PCFGs and probabilistic parsing.
Marks: 14 in the end-semester exam · CCE: Assignments or case study (12 marks, Units 3 & 4)
- N-gram language models
- Perplexity and smoothing
- Naive Bayes text classification
- Hidden Markov models
- Viterbi algorithm
- Probabilistic context-free grammars
Unit 4: NLP Tasks
9 hoursPOS tagging and named entity recognition; dependency and constituency parsing; precision, recall, F1, BLEU and ROUGE; sentiment analysis, text classification, similarity and recommendation; the vector space model; word sense disambiguation and Lesk; MFCCs in speech.
Marks: 14 in the end-semester exam · CCE: Assignments or case study (12 marks, Units 3 & 4)
- POS tagging
- Named entity recognition
- Dependency parsing
- Evaluation: BLEU and ROUGE
- Sentiment analysis
- Word sense disambiguation
Unit 5: Neural NLP and Transformers
9 hoursDense representations; attention, self-attention and scaled dot-product attention; encoder–decoder models; rule-based, statistical and neural machine translation; summarisation; multi-head attention and positional embeddings; BERT; QA and classification; chatbots and retrieval-augmented generation.
Marks: 14 in the end-semester exam · CCE: Seminar, open-book test or quiz (6 marks, Unit 5)
- Attention mechanism
- Encoder-decoder models
- Machine translation
- Transformer architecture
- BERT
- Retrieval-augmented generation
Unit contents are summarised from the official SPPU syllabus . AiCanCode.org is not affiliated with Savitribai Phule Pune University.