Track manipulation, locomotion and embodied systems.
Search papers, labs, and topics across Lattice.
Affiliation:, University of Southern California
5
0
10
8
Assigning reward credit to specific tokens, rather than uniformly, dramatically improves multi-reward RL for LLMs.
Humanoid robots can now catch thrown objects with 65% success from 2.5m using only egocentric RGB-D, a leap in reactive whole-body control.
Probabilistic GCD proposals can achieve faster convergence to target distributions with fewer particles, revolutionizing how we enforce constraints in language model outputs.
The best LLM to answer a question isn't always the best LLM to *teach* the answer, and matching the "difficulty" of the explanation to the student's current abilities yields better learning.
Current self-evolving prompt optimization frameworks falter when faced with the diverse memory extraction demands of real-world LLM assistants, but a simple clustering approach can restore generalization.