no code implementations • 5 Feb 2020 • Romina Mir, Ali Marjaninejad, Francisco J. Valero-Cuevas
Surprisingly, and contrary to recent work on manipulation, adding 1D force-sensing did not always improve learning rates compared to no sensing---likely due to whether or not normal force is relevant to the task.
1 code implementation • 10 Jul 2019 • Ali Marjaninejad, Darío Urbina-Meléndez, Francisco J. Valero-Cuevas
Error feedback is known to improve performance by correcting control signals in response to perturbations.
no code implementations • 19 Oct 2018 • Ali Marjaninejad, Darío Urbina-Meléndez, Brian A. Cohn, Francisco J. Valero-Cuevas
Robots will become ubiquitously useful only when they can use few attempts to teach themselves to perform different tasks, even with complex bodies and in dynamical environments.
1 code implementation • 15 Sep 2017 • Sébastien M. R. Arnold, Tsam Kiu Pun, Théo-Tim J. Denisart, Francisco J. Valero-Cuevas
We present Shapechanger, a library for transfer reinforcement learning specifically designed for robotic tasks.