Ulyana Piterbarg
@upiter.bsky.social
PhD at NYU studying reasoning, decision-making, and open-endedness
alum of MIT | prev: Google, MSR, MIT CoCoSci
https://upiterbarg.github.io/
alum of MIT | prev: Google, MSR, MIT CoCoSci
https://upiterbarg.github.io/
Reposted by Ulyana Piterbarg
1/13 New Paper!! We try to understand why some LMs self-improve their reasoning while others hit a wall. The key? Cognitive behaviors! Read our paper on how the right cognitive behaviors can make all the difference in a model's ability to improve with RL! 🧵
March 4, 2025 at 6:15 PM
1/13 New Paper!! We try to understand why some LMs self-improve their reasoning while others hit a wall. The key? Cognitive behaviors! Read our paper on how the right cognitive behaviors can make all the difference in a model's ability to improve with RL! 🧵
Reposted by Ulyana Piterbarg
Thank you to @sloanfoundation.bsky.social for this generous award to our lab. Hopefully this will bring us closer to building truly general-purpose robots!
🎉Congrats to the 126 early-career scientists who have been awarded a Sloan Research Fellowship this year! These exceptional scholars are drawn from 51 institutions across the US and Canada, and represent the next generation of groundbreaking researchers. sloan.org/fellowships/...
February 18, 2025 at 4:50 PM
Thank you to @sloanfoundation.bsky.social for this generous award to our lab. Hopefully this will bring us closer to building truly general-purpose robots!
Our paper showing that LMs benefit from human-like abstractions for code synthesis was accepted to ICLR! 🇸🇬
We show that order matters in code gen. -- casting code synthesis as a sequential edit problem by preprocessing examples in SFT data improves LM test-time scaling laws
We show that order matters in code gen. -- casting code synthesis as a sequential edit problem by preprocessing examples in SFT data improves LM test-time scaling laws
February 12, 2025 at 8:08 PM
Our paper showing that LMs benefit from human-like abstractions for code synthesis was accepted to ICLR! 🇸🇬
We show that order matters in code gen. -- casting code synthesis as a sequential edit problem by preprocessing examples in SFT data improves LM test-time scaling laws
We show that order matters in code gen. -- casting code synthesis as a sequential edit problem by preprocessing examples in SFT data improves LM test-time scaling laws
Reposted by Ulyana Piterbarg
Can we extend the power of world models beyond just online model-based learning? Absolutely!
We believe the true potential of world models lies in enabling agents to reason at test time.
Introducing DINO-WM: World Models on Pre-trained Visual Features for Zero-shot Planning.
We believe the true potential of world models lies in enabling agents to reason at test time.
Introducing DINO-WM: World Models on Pre-trained Visual Features for Zero-shot Planning.
January 31, 2025 at 7:24 PM
Can we extend the power of world models beyond just online model-based learning? Absolutely!
We believe the true potential of world models lies in enabling agents to reason at test time.
Introducing DINO-WM: World Models on Pre-trained Visual Features for Zero-shot Planning.
We believe the true potential of world models lies in enabling agents to reason at test time.
Introducing DINO-WM: World Models on Pre-trained Visual Features for Zero-shot Planning.
Reposted by Ulyana Piterbarg
Finally finally finally some scaling curves for imitation learning in the large-scale-data regime: arxiv.org/abs/2411.04434
Scaling Laws for Pre-training Agents and World Models
The performance of embodied agents has been shown to improve by increasing model parameters, dataset size, and compute. This has been demonstrated in domains from robotics to video games, when generat...
arxiv.org
January 20, 2025 at 2:48 PM
Finally finally finally some scaling curves for imitation learning in the large-scale-data regime: arxiv.org/abs/2411.04434
Reposted by Ulyana Piterbarg
Introducing 🧞Genie 2 🧞 - our most capable large-scale foundation world model, which can generate a diverse array of consistent worlds, playable for up to a minute. We believe Genie 2 could unlock the next wave of capabilities for embodied agents 🧠.
December 4, 2024 at 4:01 PM
Introducing 🧞Genie 2 🧞 - our most capable large-scale foundation world model, which can generate a diverse array of consistent worlds, playable for up to a minute. We believe Genie 2 could unlock the next wave of capabilities for embodied agents 🧠.
Reposted by Ulyana Piterbarg
Now that @jeffclune.bsky.social and @joelbot3000.bsky.social are here, time for an Open-Endedness starter pack.
go.bsky.app/MdVxrtD
go.bsky.app/MdVxrtD
November 20, 2024 at 7:08 AM
Now that @jeffclune.bsky.social and @joelbot3000.bsky.social are here, time for an Open-Endedness starter pack.
go.bsky.app/MdVxrtD
go.bsky.app/MdVxrtD