TY - GEN
T1 - Improving reinforcement learning with interactive feedback and affordances
AU - Cruz, Francisco
AU - Magg, Sven
AU - Weber, Cornelius
AU - Wermter, Stefan
N1 - Publisher Copyright:
© 2014 IEEE.
PY - 2014/12/11
Y1 - 2014/12/11
N2 - Interactive reinforcement learning constitutes an alternative for improving convergence speed in reinforcement learning methods. In this work, we investigate inter-agent training and present an approach for knowledge transfer in a domestic scenario where a first agent is trained by reinforcement learning and afterwards transfers selected knowledge to a second agent by instructions to achieve more efficient training. We combine this approach with action-space pruning by using knowledge on affordances and show that it significantly improves convergence speed in both classic and interactive reinforcement learning scenarios.
AB - Interactive reinforcement learning constitutes an alternative for improving convergence speed in reinforcement learning methods. In this work, we investigate inter-agent training and present an approach for knowledge transfer in a domestic scenario where a first agent is trained by reinforcement learning and afterwards transfers selected knowledge to a second agent by instructions to achieve more efficient training. We combine this approach with action-space pruning by using knowledge on affordances and show that it significantly improves convergence speed in both classic and interactive reinforcement learning scenarios.
UR - https://www.scopus.com/pages/publications/84920933638
U2 - 10.1109/DEVLRN.2014.6982975
DO - 10.1109/DEVLRN.2014.6982975
M3 - Conference contribution
AN - SCOPUS:84920933638
T3 - IEEE ICDL-EPIROB 2014 - 4th Joint IEEE International Conference on Development and Learning and on Epigenetic Robotics
SP - 165
EP - 170
BT - IEEE ICDL-EPIROB 2014 - 4th Joint IEEE International Conference on Development and Learning and on Epigenetic Robotics
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 4th Joint IEEE International Conference on Development and Learning and on Epigenetic Robotics, IEEE ICDL-EPIROB 2014
Y2 - 13 October 2014 through 16 October 2014
ER -