pouya bashivan

@PouyaBashivan

🇮🇷🇨🇦Proud husband & dad | Associate Prof @McGillPhysioloG | member @IPNMcGill @MILAMontreal | studying deep artificial/biological neural nets

Montreal, QC bashivanlab.org Joined October 2012

712 Following 1K Followers 339 Posts

Latest posts 5

Pinned post

pouya bashivan @PouyaBashivan Mar 18

Our work on building world models from episodic memories is accepted to #ICLR26!
Led by (then undergraduate, now masters student) @HeHerbie
Check out the tweeprint below!

Herbie He @HeHerbie Mar 18

New paper 🚨 #ICLR26

Most world models predict the future from a past trajectory. But neuroscience suggests that such inference can instead be made from temporally independent experiences.

We built the Episodic Spatial World Model (ESWM), a model that does exactly this:

View quoted post on X

pouya bashivan @PouyaBashivan Oct 15

I'm looking for a PhD student in my lab. Mila is one of the best academic AI research institutions with a strong neuroAI community. Please apply to if you are interested!

Mila - Institut québécois d'IA Verified account @Mila_Quebec Oct 15

Mila's annual supervision request process is now open to receive MSc and PhD applications for Fall 2027 admission! For more information, visit mila.quebec/en/prospective…

View quoted post on X

pouya bashivan @PouyaBashivan Jul 22

Do you have any fun memories from past CCN conferences? Please share your experiences and any photos you have from previous CCNs! We are organizing an event for #CCN2026 to remember previous years.

CogCompNeuro @CogCompNeuro Jul 8

CCN began at Columbia in 2017. As #CCN2026 returns to NY, a special Back2NY event will feature a crowd-sourced panel discussion on how the community has grown, prior challenges, and future perspectives.

Share your experiences in the community survey: 2026.ccneuro.org/back2ny/
View quoted post on X

pouya bashivan @PouyaBashivan Apr 24

At #ICLR2026 ? Come talk to Herbie if you are interested in world models, episodic memories, hippocampus or all of the above!

Herbie He @HeHerbie Apr 24

#ICLR2026 I’m presenting this paper today Apr 24, at pavillon 3 poster 1612, 10am-1pm. Come by if you’re interested!

Links to poster and video: iclr.cc/virtual/2026/p…
View quoted post on X

pouya bashivan @PouyaBashivan Mar 17

So proud of this work co-led by @aspa1313 and @johanobandoc
Fantastic collaboration with @AaronCourville & @pcastr
Tldr; Constraining representations within isotropic Gaussian distribution massively improves performance during RL!

Pablo Samuel Castro @pcastr Mar 17

New paper 🚨
"Stable Deep Reinforcement Learning via Isotropic Gaussian Representations"

Deep RL suffers from unstable training, representation collapse, and neuron dormancy. We show that a simple geometric insight, isotropic Gaussian representations, can fix this. Here's how 👇

View quoted post on X