
Explainability, Reasoning, Priors and GPT-3
Machine Learning Street Talk (MLST)
00:00
Intro
This chapter explores the intricacies of machine learning interpretability, contrasting inductive priors with experiential knowledge in model training. It also critiques current model explanation methods and discusses GPT-3's reasoning capabilities, emphasizing the challenges in making complex models comprehensible to the general public.
Transcript
Play full episode