Lake Presents BBT as Spectrum Between Bayesian and Neural Models
Princeton professor shares new work on combining models for better behavioral predictions.
TLDR
Brenden Lake announced the introduction of BBT. The method explores a spectrum between Bayesian and neural net models of behavior instead of treating them as separate choices. Lake suggests that combining elements from both can lead to stronger predictions of behavior. The work was developed with collaborators akjagadish and jiang_gy. It stems from material Lake presented in a keynote address. The announcement points to a video of the talk and an associated research paper.
Combined views
8.9K
1 Source, first seen ago
Sources
BL 
Brenden Lake@LakeBrendenExcited about new work from my CCN keynote (38m): https://youtu.be/v3J-vJfxhOE?si=aNHpiBCYGD9KWvCa We often choose between Bayesian and neural net models of behavior. But what if there's a spectrum, and combining makes better predictions? Introducing BBT w @akjagadish @jiang_gy…
- likes: 88
- replies: 3
- bookmarks: 36
- reposts: 11