{"id":1380,"date":"2023-07-17T09:10:36","date_gmt":"2023-07-17T08:10:36","guid":{"rendered":"https:\/\/babel.isa.uma.es\/kipr\/?p=1380"},"modified":"2023-07-17T10:05:03","modified_gmt":"2023-07-17T09:05:03","slug":"a-nice-survey-on-active-learning-in-particular-for-robotics","status":"publish","type":"post","link":"https:\/\/babel.isa.uma.es\/kipr\/?p=1380","title":{"rendered":"A nice survey on active learning, in particular for robotics"},"content":{"rendered":"

Annalisa T. Taylor, Thomas A. Berrueta, Todd D. Murphey, Active learning in robotics: A review of control principles,<\/strong> . Mechatronics, Volume 77, 2021 DOI: 10.1016\/j.mechatronics.2021.102576<\/a>.<\/h4>\n

Active learning is a decision-making process. In both abstract and physical settings, active learning demands
\nboth analysis and action. This is a review of active learning in robotics, focusing on methods amenable to
\nthe demands of embodied learning systems. Robots must be able to learn efficiently and flexibly through
\ncontinuous online deployment. This poses a distinct set of control-oriented challenges??one must choose
\nsuitable measures as objectives, synthesize real-time control, and produce analyses that guarantee performance
\nand safety with limited knowledge of the environment or robot itself. In this work, we survey the fundamental
\ncomponents of robotic active learning systems. We discuss classes of learning tasks that robots typically
\nencounter, measures with which they gauge the information content of observations, and algorithms for
\ngenerating action plans. Moreover, we provide a variety of examples ?? from environmental mapping to
\nnonparametric shape estimation ?? that highlight the qualitative differences between learning tasks, information
\nmeasures, and control techniques. We conclude with a discussion of control-oriented open challenges, including
\nsafety-constrained learning and distributed learning.<\/p><\/blockquote>\n

NOTES:<\/strong><\/p>\n