Zobrazeno 1 - 10
of 128
pro vyhledávání: '"Zhang Jesse"'
Publikováno v:
CoRL 2024
Most reinforcement learning (RL) methods focus on learning optimal policies over low-level action spaces. While these methods can perform well in their training environments, they lack the flexibility to transfer to new tasks. Instead, RL agents that
Externí odkaz:
http://arxiv.org/abs/2406.17768
Autor:
Wang, Yufei, Sun, Zhanyi, Zhang, Jesse, Xian, Zhou, Biyik, Erdem, Held, David, Erickson, Zackory
Reward engineering has long been a challenge in Reinforcement Learning (RL) research, as it often requires extensive human effort and iterative processes of trial-and-error to design effective reward functions. In this paper, we propose RL-VLM-F, a m
Externí odkaz:
http://arxiv.org/abs/2402.03681
We propose a framework that leverages foundation models as teachers, guiding a reinforcement learning agent to acquire semantically meaningful behavior without human feedback. In our framework, the agent receives task instructions grounded in a train
Externí odkaz:
http://arxiv.org/abs/2312.08958
Autor:
Zhang, Jesse, Zhang, Jiahui, Pertsch, Karl, Liu, Ziyi, Ren, Xiang, Chang, Minsuk, Sun, Shao-Hua, Lim, Joseph J.
We propose BOSS, an approach that automatically learns to solve new long-horizon, complex, and meaningful tasks by growing a learned skill library with minimal supervision. Prior work in reinforcement learning require expert supervision, in the form
Externí odkaz:
http://arxiv.org/abs/2310.10021
Autor:
Sontakke, Sumedh A, Zhang, Jesse, Arnold, Sébastien M. R., Pertsch, Karl, Bıyık, Erdem, Sadigh, Dorsa, Finn, Chelsea, Itti, Laurent
Reward specification is a notoriously difficult problem in reinforcement learning, requiring extensive expert supervision to design robust reward functions. Imitation learning (IL) methods attempt to circumvent these problems by utilizing expert demo
Externí odkaz:
http://arxiv.org/abs/2310.07899
Autor:
Liu, Zuxin, Zhang, Jesse, Asadi, Kavosh, Liu, Yao, Zhao, Ding, Sabach, Shoham, Fakoor, Rasool
The full potential of large pretrained models remains largely untapped in control domains like robotics. This is mainly because of the scarcity of data and the computational challenges associated with training or fine-tuning these large models for su
Externí odkaz:
http://arxiv.org/abs/2310.05905
Pre-training robot policies with a rich set of skills can substantially accelerate the learning of downstream tasks. Prior works have defined pre-training tasks via natural language instructions, but doing so requires tedious human annotation of hund
Externí odkaz:
http://arxiv.org/abs/2306.11886
Program synthesis aims to automatically construct human-readable programs that satisfy given task specifications, such as input/output pairs or demonstrations. Recent works have demonstrated encouraging results in a variety of domains, such as string
Externí odkaz:
http://arxiv.org/abs/2303.06018
Recently, deep reinforcement learning (DRL) methods have achieved impressive performance on tasks in a variety of domains. However, neural network policies produced with DRL methods are not human-interpretable and often have difficulty generalizing t
Externí odkaz:
http://arxiv.org/abs/2108.13643
Autor:
Burkley, Zakary, Borges, Lucas de Sousa, Ohayon, Ben, Golovozin, Artem, Zhang, Jesse, Crivelli, Paolo
We demonstrate the superior performance of fluoride coated versus oxide coated mirrors in long term vacuum operation of a high power deep-ultraviolet enhancement cavity. In high vacuum ($10^{-8}$ mbar), the fluoride optics can maintain up to a record
Externí odkaz:
http://arxiv.org/abs/2105.14977