Shahab Bakhtiari
@shahabbakht.bsky.social did:plc:dpgtuy5vtitbaayhzhe4mctb
|| assistant prof at University of Montreal || leading the systems neuroscience and AI lab (SNAIL: https://www.snailab.ca/) 🐌 || associate academic member of Mila (Quebec AI Institute) || #NeuroAI || vision and learning in brains and machines
Sur Bluesky depuis le
- Abonnés
- 6 585
- Abonnements
- 1 184
- Posts
- 1 567
- Jours dans les classements
- 0
Résumé
Shahab Bakhtiari (@shahabbakht.bsky.social) est sur Bluesky depuis le 5 août 2023. Le compte a 6 585 abonnés, suit 1 184 comptes et a publié 1 567 posts.
Présence dans les classements
Les meilleurs jours où les posts de ce compte sont entrés dans le top quotidien de Debya.
Ce compte n’est encore apparu dans aucun classement quotidien.
Posts aimés récemment
J’aime publics du réseau Bluesky.
- Nancy Kanwisher@nancykanwisher.bsky.social
New paper with Celia Bougou, Richard Andersen, Josh Tenenbaum, @rtpramod.bsky.social and others: A Unified World Model in Human Posterior Parietal Cortex. Shoutout to heroic participant RD who let us record from individual neurons while he performed our tasks. doi.org/10.64898/202... A Thread...
Voir le post - Brad Wyble@bwyble.bsky.social
AI has not made theory any cheaper. AI has made it easier to develop code and tools that allow us to test things, somewhat lowering the bar for testing some kinds of theories.
Voir le post 👀An important one for representational “stability-flexibility“ and drift debaters, including PV INs regulating drift, separable from task. Yet another noteworthy finding from Wenbo Tang in Oliva and @antferrui.bsky.social labs 🧪 🧠📈 www.nature.com/articles/s41...
Voir le postNew theory about value computations in PFC from the awesome @jowarrn.bsky.social and @jcrwhittington.bsky.social V excited that we can start thinking about sequence/schema tasks and value tasks in the same language. (This time with working videos:)
Voir le postIs meta-RL in PFC just a big RNN or can we figure out its mechanism? Here we show that RNNs learn a particular, but general, method for meta-RL. They place options in 'ranked’ subspaces (slots), and dynamically re-rank them. Value doesn’t pick the action; it reorders the slots 🧵 (Reposted w/ vids)
Voir le post
Comptes bloqués
Sur Bluesky, les blocages sont publics.
Ce compte ne bloque personne.