Streaming Reinforcement Learning (RL) is a huge challenge: transitions are used once and discarded immediately. This makes agents extremely sample-inefficient. But what if we could "squeeze" more information out of every single frame?

Check out our latest paper!

New work, just accepted @ICLR: "The Expressive Limits of Diagonal SSMs for State-Tracking"
We give a complete characterization of what diagonal SSMs can and cannot compute on state-tracking tasks and the answer is deeply connected to group theory.
🧵👇

Can LLMs play Hangman? Spoiler alert: Not yet.
Check out “LLMs Can’t Play Hangman: On the Necessity of a Private Working Memory for Language Agents”, led by Davide Baldelli, Ali Parviz, AmalZouaq and Sarath Chandar.

Can LLMs become CAD designers?
Check out “CADmium: Fine-Tuning Code Language Models for Text-Driven Sequential CAD Design”, which is now published in Transactions on Machine Learning Research (TMLR)!

Life update - last month I moved to #montreal 🇨🇦 from #Sydney 🇦🇺 to kick off my @ivado.bsky.social postdoc fellowship at @mila-quebec.bsky.social. Must say I am constantly amused by: 1. How walkable the city is.
2. How easy is it to reach out to diverse research communities within #mila ! 😀

🎉 Happy to share that our paper “Mining your own secrets: Diffusion Classifier scores for Continual Personalization of Text-to-Image Diffusion Models” has been accepted to #ICLR2025!

👉 The work results from my #Sony summer internship in the stunning #Tokyo🗼 city

Preprint: arxiv.org/pdf/2410.00700

I ran across a busy Sander at a #neurips party with a similar question - he was still patient enough to explain stuff. This talk further clarifies a good amount of my doubts. Recommend watching if you're working on diffusion / LLMs for generation!

I validate this