Yoav Ger

Today at the last poster [3-158] session of #cosyne2026 come here about 3 works from the Barak lab.

@kabirdabholkar.bsky.social @cosynemeeting.bsky.social

This was a fantastic poster presentation!

At #NeurIPS? Curious about how RNNs learn differently in closed-loop (RL) vs. open-loop (supervised) settings?
Come by Poster #2107 on Thursday at 4:30 PM!

neurips.cc/virtual/2025...

#NeuRIPS2025
Wanna find decision boundaries in your RNN? Or learn about Koopman Eigenfunctions? Come to my poster.

#2002
Wednesday, Dec 3, 11am-2pm.
Exhb. Hall C,D,E
San Diego

neurips.cc/virtual/2025...

Are world models necessary to achieve human-level agents, or is there a model-free short-cut?
Our new #ICML2025 paper tackles this question from first principles, and finds a surprising answer, agents _are_ world models… 🧵
arxiv.org/abs/2506.01622

Learning Dynamics of RNNs in Closed-Loop Environments Recurrent neural networks (RNNs) trained on neuroscience-inspired tasks offer powerful models of brain computation. However, typical training paradigms rely on open-loop, supervised settings, whereas ...

All our motor control modelling efforts focus on closed-loop systems for this reason:

"...closed-loop and open-loop training produce fundamentally different learning dynamics, even when using identical architectures and converging to the same final solution."

arxiv.org/abs/2505.13567

מדויק