Machine Learning Street Talk (MLST) DR. JEFF BECK - THE BAYESIAN BRAIN
49 snips
Oct 16, 2023 
Dr. Jeff Beck, a computational neuroscientist, dives into the fascinating world of probabilistic reasoning in both humans and animals. He discusses the Bayesian brain hypothesis and its impact on understanding decision-making under uncertainty. Active inference takes center stage as Beck explains how agents interact with their environments. The conversation also explores the philosophy behind consciousness, the integration of cognition in ecosystems, and innovations in Bayesian multi-agent systems. Expect a thought-provoking journey blending science and philosophy!
AI Snips
Chapters
Transcript
Episode notes
Bayesian Models and Empirical Inquiry
- Bayesian models are fundamental to empirical inquiry, offering a normative approach.
- The best approach involves considering all possible models and letting Bayesian inference guide selection.
Conditional Knowledge and Model Selection
- All knowledge is conditional, relying on acquired data and considered models.
- Ideally, consider all possible models, but practically, focus on computationally accessible ones.
Evolutionary Priors and Inference
- Humans possess an elaborate set of priors refined by evolution.
- These priors allow for efficient inference but don't necessitate de novo reasoning.