Learning options for an MDP from demonstrations

Tamassia, M, Zambetta, F, Raffe, W and Li, X 2015, 'Learning options for an MDP from demonstrations', in Stephan K. Chalup, Alan D. Blair, and Marcus Randall (ed.) Artificial Life and Computational Intelligence, Newcastle, Australia, 5-7 February 2015, pp. 226-242.

Document type: Conference Paper
Collection: Conference Papers

Title Learning options for an MDP from demonstrations
Author(s) Tamassia, M
Zambetta, F
Raffe, W
Li, X
Year 2015
Conference name First Australasian Conference on Artificial Life and Computational Intelligence
Conference location Newcastle, Australia
Conference dates 5-7 February 2015
Proceedings title Artificial Life and Computational Intelligence
Editor(s) Stephan K. Chalup, Alan D. Blair, and Marcus Randall
Publisher Springer International Publishing
Place of publication Switzerland
Start page 226
End page 242
Total pages 17
Abstract The options framework provides a foundation to use hierarchical actions in reinforcement learning. An agent using options, along with primitive actions, at any point in time can decide to perform a macro-action made out of many primitive actions rather than a primitive action. Such macro-actions can be hand-crafted or learned. There has been previous work on learning them by exploring the environment. Here we take a different perspective and present an approach to learn options from a set of experts demonstrations. Empirical results are also presented in a similar setting to the one used in other works in this area.
Subjects Adaptive Agents and Intelligent Robotics
Keyword(s) agent-based modeling - artificial neural network - autonomous agents - case-based reasoning - classification - computational intelligence - developmental robotics - dynamical systems - evolutionary algorithm - genetic algorithms - image retrieval - machine learning - machine translation - multi-agent systems - ontology - self organization - simulation - support vector machines - topic modeling - virtual agents
Copyright notice © 2015 Springer International Publishing Switzerland
ISSN 0302-9743
Version Filter Type
Access Statistics: 154 Abstract Views  -  Detailed Statistics
Created: Mon, 20 Apr 2015, 11:57:00 EST by Catalyst Administrator
© 2014 RMIT Research Repository • Powered by Fez SoftwareContact us