Deep COACH: Interactive Learning with Corrective Feedback for Policies based on Deep Neural Networks

Описание к видео Deep COACH: Interactive Learning with Corrective Feedback for Policies based on Deep Neural Networks

Learning deep neural networks policies interactively based on corrective human feedback. In the video there are examples of learning with human corrections, which were carried out after training an Auto-Encoder with data recorded from demonstrations.

"Interactive Learning with Corrective Feedback for Policies based on Deep Neural Networks"
Pérez-Dattari, Rodrigo; Celemin, Carlos; Ruiz-del-Solar, Javier; Kober, Jens
International Symposium on Experimental Robotics (ISER) 2018

Комментарии

Информация по комментариям в разработке