IIT Madras NLP Lectures: Introduction to Natural Language Processing
BSDA500561 lectures12 weeks
Weeks
- Week 14 lecturesUnderstanding natural language and processing · History of nlp · Why it is hard to process natural language? · Levels of language processing and linguistics fundamentals
- Week 24 lecturesText processing · Text processing demo · Text analysis – stemming and lemmatization · Morphological analysis
- Week 36 lecturesPart of speech tagging · Methods of parts of speech tagging · Named entity recognition · Sequence modeling – HMM, MEMM, and CRF · Classification models – naïve bayes, logistic regression, clustering · Demo of HMM and CRF
- Week 45 lecturesSyntax, context-free grammar, and constituency parsing · Dependency parsing · Meaning representation – logical semantics & semantic role labelling · Transition graph-based parser – code demo · Semantic role labelling – code demo
- Week 55 lecturesDistributional semantics, n-gram, word2vec, and glove · Anaphora and coreference resolution, discourse connectives · Machine translation · Word2vec and glove – code demo · Machine translation – code demo
- Week 64 lecturesRecurrent neural networks · Pytorch installation and basic overview · Data preprocessing for nlp tasks · Long short-term memory (LSTM) implementation
- Week 75 lecturesTransformers – part 1 · Transformers – part 2 · Large language models – part 1 · Large language models – part 2 · Transformers implementation
- Week 86 lecturesLow-rank adaptation (lora) · QLoRA (Quantized Low-Rank Adaptation) · Intuition for instruction fine-tuning · Supervised instruction fine-tuning · Human alignment, rlhf, and dpo · Large language models (LLMs) with hugging face
- Week 96 lecturesIntroduction to nlg · Decoding in nlg · Beam search decoding · Sampling-based decoding · Speculative sampling · Fine-tuning an large language models using unsloth
- Week 106 lecturesMotivation for retrieval-augmented generation · Retrieval-augmented generation · Application – question answering · Application – code optimization · Application – text summarization · Retrieval augmented generation (RAG) and applications
- Week 115 lecturesMotivation for explainability · Attention-based explainability · Attribution-based methods · Examples of explainability · Model explainability using integrated gradients and attention maps
- Week 125 lecturesEthical considerations in nlp · Bias – background and seminal debiasing methods · Bias in LLMs and metrics for evaluation · Bias mitigation techniques · Bias – additional risk factors and recommendations for utilizing LLMs in research