ICML '06 Proceedings of the 23rd international conference on Machine learning
Active Learning for Reward Estimation in Inverse Reinforcement Learning
ECML PKDD '09 Proceedings of the European Conference on Machine Learning and Knowledge Discovery in Databases: Part II
Convex and Semi-Nonnegative Matrix Factorizations
IEEE Transactions on Pattern Analysis and Machine Intelligence
Basis Decomposition of Motion Trajectories Using Spatio-temporal NMF
ICANN '09 Proceedings of the 19th International Conference on Artificial Neural Networks: Part II
The facilitatory role of linguistic instructions on developing manipulation skills
IEEE Computational Intelligence Magazine
Autonomous Helicopter Aerobatics through Apprenticeship Learning
International Journal of Robotics Research
Human behavior understanding for robotics
HBU'12 Proceedings of the Third international conference on Human Behavior Understanding
Socially guided intrinsic motivation for robot learning of motor skills
Autonomous Robots
| Hi-index | 0.00 |
In many applications, such as virtual agents or humanoid robots, it is difficult to represent complex human behaviors and the full range of skills necessary to achieve them. Real life human behaviors are often the combination of several parts and never reproduced in the exact same way. In this work we introduce a new algorithm that is able to learn behaviors by assuming that the observed complex motions can be represented in a smaller dictionary of concurrent tasks. We present an optimization formalism and show how we can learn simultaneously the dictionary and the mixture coefficients that represent each demonstration. We present results on a idealized model where a set of potential functions represents human objectives or preferences for achieving a task.