Finetuning and Adaptation

Chapter 27: Finetuning Techniques Full finetuning, parameter-efficient tuning Adapters, LoRA, prefix tuning [PEFT (Parameter-Efficient Fine-Tuning), hyperparameter tuning, catastrophic forgetting] References Chapter 28: Domain-Specific NLP Domain adaptation, specialized corpora Applications: Medical, legal NLP [Domain-adaptive pretraining, BioBERT, LegalBERT] References Chapter 29: Few-Shot Learning Prompt engineering, in-context learning Applications: T5, GPT-3 [Zero-shot learning, meta-learning, prompt design frameworks] References