Constructing skill trees for reinforcement learning agents from demonstration trajectories
Publication
, Journal Article
Konidaris, G; Kuindersmay, S; Barto, A; Grupen, R
Published in: Advances in Neural Information Processing Systems 23: 24th Annual Conference on Neural Information Processing Systems 2010, NIPS 2010
December 1, 2010
We introduce CST, an algorithm for constructing skill trees from demonstration trajectories in continuous reinforcement learning domains. CST uses a changepoint detection method to segment each trajectory into a skill chain by detecting a change of appropriate abstraction, or that a segment is too complex to model as a single skill. The skill chains from each trajectory are then merged to form a skill tree. We demonstrate that CST constructs an appropriate skill tree that can be further refined through learning in a challenging continuous domain, and that it can be used to segment demonstration trajectories on a mobile manipulator into chains of skills where each skill is assigned an appropriate abstraction.
Duke Scholars
Published In
Advances in Neural Information Processing Systems 23: 24th Annual Conference on Neural Information Processing Systems 2010, NIPS 2010
Publication Date
December 1, 2010
Citation
APA
Chicago
ICMJE
MLA
NLM
Konidaris, G., Kuindersmay, S., Barto, A., & Grupen, R. (2010). Constructing skill trees for reinforcement learning agents from demonstration trajectories. Advances in Neural Information Processing Systems 23: 24th Annual Conference on Neural Information Processing Systems 2010, NIPS 2010.
Konidaris, G., S. Kuindersmay, A. Barto, and R. Grupen. “Constructing skill trees for reinforcement learning agents from demonstration trajectories.” Advances in Neural Information Processing Systems 23: 24th Annual Conference on Neural Information Processing Systems 2010, NIPS 2010, December 1, 2010.
Konidaris G, Kuindersmay S, Barto A, Grupen R. Constructing skill trees for reinforcement learning agents from demonstration trajectories. Advances in Neural Information Processing Systems 23: 24th Annual Conference on Neural Information Processing Systems 2010, NIPS 2010. 2010 Dec 1;
Konidaris, G., et al. “Constructing skill trees for reinforcement learning agents from demonstration trajectories.” Advances in Neural Information Processing Systems 23: 24th Annual Conference on Neural Information Processing Systems 2010, NIPS 2010, Dec. 2010.
Konidaris G, Kuindersmay S, Barto A, Grupen R. Constructing skill trees for reinforcement learning agents from demonstration trajectories. Advances in Neural Information Processing Systems 23: 24th Annual Conference on Neural Information Processing Systems 2010, NIPS 2010. 2010 Dec 1;
Published In
Advances in Neural Information Processing Systems 23: 24th Annual Conference on Neural Information Processing Systems 2010, NIPS 2010
Publication Date
December 1, 2010