Výsledky vyhledávání

Report

Meta-Designing Quantum Experiments with Language Models

Autor: Arlt, Sören, Duan, Haonan, Li, Felix, Xie, Sang Michael, Wu, Yuhuai, Krenn, Mario

Artificial Intelligence (AI) has the potential to significantly advance scientific discovery by finding solutions beyond human capabilities. However, these super-human solutions are often unintuitive and require considerable effort to uncover underly

Externí odkaz: http://arxiv.org/abs/2406.02470

Zobrazit plný text záznamu

Dissertation/ Thesis

Wok Palace 88

Autor: Li, Felix

“Wok Palace 88,” is a parafictional and defunct Chinese-American restaurant brought back to life through the institutional process of excavating, archiving, and displaying. The restaurant is based around shutdown restaurants in Greater Boston inc

Externí odkaz: https://hdl.handle.net/1721.1/151372

Zobrazit plný text záznamu

Report

New Parameter Measurements for the Ultra-Short-Period Planet TOI-1807b

Autor: Peng, Peifeng, Xiong, Hongyi, Li, He, Li, Felix, Wang, Tianqi

The ultra-short-period (USP) planets are exoplanets with very short orbital periods ($\textit{P} < 1$ day), and TOI-1807b is one such planet recently discovered by the TESS mission where it orbits in the TOI-1807 system that is still little known now

Externí odkaz: http://arxiv.org/abs/2210.04162

Zobrazit plný text záznamu

Report

Insights into Pre-training via Simpler Synthetic Tasks

Autor: Wu, Yuhuai, Li, Felix, Liang, Percy

Pre-training produces representations that are effective for a wide range of downstream tasks, but it is still unclear what properties of pre-training are necessary for effective gains. Notably, recent work shows that even pre-training on synthetic t

Externí odkaz: http://arxiv.org/abs/2206.10139

Zobrazit plný text záznamu

Report

On-Policy Robot Imitation Learning from a Converging Supervisor

Autor: Balakrishna, Ashwin, Thananjeyan, Brijen, Lee, Jonathan, Li, Felix, Zahed, Arsh, Gonzalez, Joseph E., Goldberg, Ken

Publikováno v: 3rd Conference on Robot Learning (CoRL 2019)

Existing on-policy imitation learning algorithms, such as DAgger, assume access to a fixed supervisor. However, there are many settings where the supervisor may evolve during policy learning, such as a human performing a novel task or an improving al

Externí odkaz: http://arxiv.org/abs/1907.03423

Zobrazit plný text záznamu

Report

Safety Augmented Value Estimation from Demonstrations (SAVED): Safe Deep Model-Based RL for Sparse Cost Robotic Tasks

Autor: Thananjeyan, Brijen, Balakrishna, Ashwin, Rosolia, Ugo, Li, Felix, McAllister, Rowan, Gonzalez, Joseph E., Levine, Sergey, Borrelli, Francesco, Goldberg, Ken

Publikováno v: Robotics and Automation Letters 2020

Reinforcement learning (RL) for robotics is challenging due to the difficulty in hand-engineering a dense cost function, which can lead to unintended behavior, and dynamical uncertainty, which makes exploration and constraint satisfaction challenging

Externí odkaz: http://arxiv.org/abs/1905.13402

Zobrazit plný text záznamu