Advanced NLP: Text-based QA

Explore text-based question answering in this advanced Natural Language Processing lecture from Carnegie Mellon University. Delve into machine reading datasets, context encoding methods, and multi-hop reasoning techniques. Examine various question types including multiple choice, span selection, and closed questions. Learn about word classification, generative QA models, and retrieval-based approaches such as Locality Sensitive Passing and Token Level Retrieval. Investigate Retrieval Augmented Generation and Mass Language Models. Analyze datasets like Baby Data Set and CNN Daily Mail, while considering model distraction and adversarial creation. Gain insights into the caveats of datasets and the challenges in developing robust QA systems.

Introduction
Textbased QA
Questions
Multiple Choice Questions
Span Selection
Closed Questions
generative QA
models for machine reading
word classification
generative QA models
Retrievalbased QA
Locality Sensitive Passing
Token Level Retrieval
Retrieval Augmented Generation
Mass Language Models
Data Sets
Baby Data Set
CNN Daily Mail
Model Distraction
Adversarial Creation