BSDA5005 — Natural Language Processing (NLP)
339 words
2 min read
Visual companion
Python
Type and operator map
Python Week 1: the first filter for runtime behavior
View
Revision summary
What this note is really saying
Short form
# BSDA5005 — Natural Language Processing (NLP) > **Course**: Degree Elective (4 credits) **Topics**: Text preprocessing, POS tagging, NER, parsing, word embeddings, seq2seq, transformers for NLP **Files**: 12 in-depth topic files across 12 weeks ## Course Index ### Week 1: Introduction & Text Preprocessing 1. **[NLP...
BSDA5005 — Natural Language Processing (NLP)
Course: Degree Elective (4 credits) Topics: Text preprocessing, POS tagging, NER, parsing, word embeddings, seq2seq, transformers for NLP Files: 12 in-depth topic files across 12 weeks
Course Index
Week 1: Introduction & Text Preprocessing
- NLP Introduction & Text Preprocessing — Linguistic levels, tokenization, normalization, stemming, lemmatization
Week 2: POS Tagging
- Part-of-Speech Tagging — Stochastic tagging, HMM, Viterbi algorithm, tag transitions
Week 3: Parsing & Syntax
- Parsing & Syntactic Analysis — Constituency parsing, dependency parsing, CFG, CYK algorithm
Week 4: Named Entity Recognition
- Named Entity Recognition — Sequence labeling, BIO tagging, CRF, feature engineering
Week 5: Word Embeddings
- Word Embeddings: Word2Vec & GloVe — Distributional semantics, CBOW, Skip-gram, GloVe, FastText
Week 6: Text Classification
- Text Classification & TF-IDF — TF-IDF, Naive Bayes, logistic regression, evaluation metrics
Week 7: Sequence Models
- RNNs & LSTMs for NLP — Sequence modeling, LSTM gates, bidirectional RNNs
Week 8: Sequence-to-Sequence
- Seq2Seq & Attention — Encoder-decoder, attention mechanism, teacher forcing, beam search
Week 9: Sentiment Analysis
- Sentiment Analysis & Opinion Mining — Polarity, aspect-based, lexicon methods, deep learning approaches
Week 10: Machine Translation
- Machine Translation & Evaluation — Statistical MT, neural MT, BLEU, ROUGE, evaluation challenges
Week 11: Contextual Embeddings
- ELMo, BERT & Pre-trained Models — Contextual vs static embeddings, transfer learning in NLP
Week 12: Advanced NLP
- Advanced Topics: Transformers, Prompting & Beyond — Transformer architectures for NLP, few-shot learning, NLP pipeline design
Key Concepts
| Concept | Description |
|---|---|
| Tokenization | Splitting text into tokens (words, subwords) |
| POS Tagging | Assigning grammar tags (NN, VB, JJ) to words |
| NER | Identifying entities (Person, Location, Organization) |
| Parsing | Analyzing grammatical structure |
| Word Embeddings | Dense vector representations of words |
| Seq2Seq | Sequence-to-sequence modeling with attention |
| BLEU/ROUGE | Evaluation metrics for generation |
Exam Weightage
| Topic | Quiz 1 | Quiz 2 | End Term |
|---|---|---|---|
| Linguistic Fundamentals | ★★★★★ | ★★★ | ★★★ |
| POS Tagging | ★★★★★ | ★★ | ★★★★ |
| Parsing | ★★★★ | ★★ | ★★★★ |
| NER | ★★★ | ★★★ | ★★★ |
| Word Embeddings | ★★ | ★★★★★ | ★★★ |
| Text Classification | ★★ | ★★★★ | ★★★ |
| Sequence Models | ★ | ★★★★★ | ★★★★ |
| Machine Translation | ★ | ★★★ | ★★★★★ |
| Evaluation Metrics | ★★ | ★★★ | ★★★★★ |
Cross-Course Links
- BSDA5004 (LLMs): Tokenization, attention, pre-training concepts
- BSDA5002 (GenAI Foundations): Probability and information theory foundations
- BSDA5013 (DL Practice): Hands-on NLP implementation projects Join Discord PreviousEvaluation Metrics