Abstract
Imitation learning is a promising route to instruct robotic multi-agent systems. However, imitating agents should be able to decide autonomously what behavior, observed in others, is interesting to copy. Here we investigate whether a simple recurrent network (Elman net) can be used to extract meaningful chunks from a continuous sequence of observed actions. Results suggest that, even in spite of the high level of task specific noise, Elman nets can be used for isolating re-occurring action patterns in robots. Limitations and future directions are discussed.
Original language | English |
---|---|
Title of host publication | Intelligent Systems 2008 : IS '08 ; 4th International IEEE Conference, 6-8 September, 2008, Varna, Bulgaria |
Place of Publication | Piscataway |
Publisher | Institute of Electrical and Electronics Engineers |
Pages | 10-7-10-13 |
Volume | 2 |
ISBN (Print) | 978-1-4244-1739-1 |
DOIs | |
Publication status | Published - 2008 |
Event | conference; IS '08 : Intelligent Systems 2008 ; 4th International IEEE Conference, 6-8 September, 2008, Varna, Bulgaria - Duration: 1 Jan 2008 → … |
Conference
Conference | conference; IS '08 : Intelligent Systems 2008 ; 4th International IEEE Conference, 6-8 September, 2008, Varna, Bulgaria |
---|---|
Period | 1/01/08 → … |
Other | IS '08 : Intelligent Systems 2008 ; 4th International IEEE Conference, 6-8 September, 2008, Varna, Bulgaria |