Thomas Rupf (@th_rupf) profile picture

Thomas Rupf

@th_rupf

Master's student @ ETH interested in ML, RL, and Robotics.

 Zurich
Joined July 2025
  • Tweets12
  • Following169
  • Followers23
If you're at ICLR馃嚙馃嚪 come check out OpTI-BFM at poster P4-4504 tomorrow morning or as an oral at 4:03pm in the Amphitheater.
Thomas Rupf (@th_rupf) profile pictureThomas Rupf@th_rupfFeb 5
Excited to share that our paper "Optimistic Task Inference for Behavior Foundation Models" was accepted for ICLR 2026. BFMs are great at zero-shot RL, but task inference requires a dataset with reward labels. Our method OpTI-BFM offers an online alternative. (1/5)
Image from Thomas Rupf (@th_rupf)
1
4
16
1,008
Excited to share that our paper "Optimistic Task Inference for Behavior Foundation Models" was accepted for ICLR 2026. BFMs are great at zero-shot RL, but task inference requires a dataset with reward labels. Our method OpTI-BFM offers an online alternative. (1/5)
Image from Thomas Rupf (@th_rupf)
1
5
2
12
2,221
OpTI-BFM infers the task online through interaction with the environment. It uses the observed rewards to update its uncertainty over the task embeddings and then chooses the most optimistic one to collect informative data. (2/5)
1
61
Zero-shot imitation from just a single sparse demonstration is hard. Goal-conditioned methods tend to 鈥済reedily" move from one state to the next and lose the big picture. We're presenting an alternative approach on Tuesday at #ICML2025. (1/3)
1
7
16
1,616
Our method tackles the occupancy matching objective directly at test-time by estimating the agent's occupancy with samples from a learned world model and matching it to the expert occupancy using Optimal Transport. (2/3)
Image from Thomas Rupf (@th_rupf)
1
1
72
Load more