Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Accelerating imitation learning in relational domains via transfer by initialization

: Natarajan, S.; Odom, P.; Joshi, S.; Khot, T.; Kersting, K.; Tadepalli, P.


Zaverucha, G.:
Inductive logic programming. 23rd International Conference, ILP 2013 : Rio de Janeiro, Brazil, August 28-30, 2013; Revised selected papers
Berlin: Springer, 2014 (Lecture Notes in Computer Science 8812)
ISBN: 978-3-662-44922-6 (Print)
ISBN: 978-3-662-44923-3 (Online)
ISBN: 3-662-44922-6
International Conference on Inductive Logic Programming (ILP) <23, 2013, Rio de Janeiro>
Fraunhofer IAIS ()

The problem of learning to mimic a human expert/teacher from training trajectories is called imitation learning. Tomake the process of teaching easier in this setting, we propose to employ transfer learning (where one learns on a source problem and transfers the knowledge to potentially more complex target problems). We consider multi-relational environments such as real-time strategy games and use functional-gradient boosting to capture and transfer the models learned in these environments. Our experiments demonstrate that our learner learns a very good initial model from the simple scenario and effectively transfers the knowledge to the more complex scenario thus achieving a jump start, a steeper learning curve and a higher convergence in performance.