@rmounir discusses unsupervised learning, shared learning in columns, and voting as an inference process. The team explores challenges in learning behavior models, object IDs, and how columns might collaborate to build complete representations from partial observations.
Summary Video
Main Video
0:00 Introduction
5:37 Unsupervised Learning
13:50 Learning Behavior
17:01 Inferring Behavior
27:01 Shared Learning for Behaviors
41:49 Voting is an Inference Process
54:53 The Learning of Object IDs
1:20:19 Generalized Pooler in Hierarchy
1:24:25 Discussion: The High-Level Idea
1:55:05 Recurrent Projections in the Neocortex
1:59:07 Generalized Pooler
2:22:45 Further Discussions