An Active Learning Framework for Efficient Robust Policy Search

• Published in: arXiv.org
• Main Authors: Narayanaswami, Sai Kiran, Nandan Sudarsanam, Ravindran, Balaraman
• Format: Paper Journal Article
• Language: English
• Published: Ithaca Cornell University Library, arXiv.org 21.11.2021
• Subjects: Active learning; Computer Science - Artificial Intelligence; Computer Science - Learning; Computer Science - Robotics; Computer simulation; Control tasks; Environment models; Learning; Mathematical models; Parameters; Performance degradation; Policies; Robustness; Searching; Statistics - Machine Learning

Robust Policy Search is the problem of learning policies that do not degrade in performance when subject to unseen environment model parameters. It is particularly relevant for transferring policies learned in a simulation environment to the real world. Several existing approaches involve sampling l...