Abstract: In this work, we address the problem of motion planning for control-affine systems. We first employ the Geometric Heat Flow (GHF) method to generate an admissible trajectory that adheres to ...

Abstract: We consider the problem of learning the dynamics of a linear system when one has access to data generated by an auxiliary system that shares similar (but ...

REC-R1 is a general framework that bridges generative large language models (LLMs) and recommendation systems via reinforcement learning. Check the paper here.

KINESIS is a model-free imitation-learning framework that facilitates the development of effective and scalable muscle-based control policies of locomotion. KINESIS is trained on 1.8 hours of ...