AGI 2016: Artificial General Intelligence pp 64-73 | Cite as
Imitation Learning as Cause-Effect Reasoning
Conference paper
First Online:
Abstract
We propose a framework for general-purpose imitation learning centered on cause-effect reasoning. Our approach infers a hierarchical representation of a demonstrator’s intentions, which can explain why they acted as they did. This enables rapid generalization of the observed actions to new situations. We employ a novel causal inference algorithm with formal guarantees and connections to automated planning. Our approach is implemented and validated empirically using a physical robot, which successfully generalizes skills involving bimanual manipulation of composite objects in 3D. These results suggest that cause-effect reasoning is an effective unifying principle for cognitive-level imitation learning.
Keywords
Artificial general intelligence Imitation learning Cause-effect reasoning Parsimonious covering theory Cognitive roboticsNotes
Acknowledgements
This work was supported by ONR award N000141310597. Thanks to Ethan Reggia for building the hard-drive docking station.
References
- 1.Akgun, B., Cakmak, M., Yoo, J.W., Thomaz, A.L.: Trajectories and keyframes for kinesthetic teaching. In: Proceedings of the 7th Annual ACM/IEEE International Conference on Human-Robot Interaction, pp. 391–398. ACM (2012)Google Scholar
- 2.Baldwin, D., Baird, J.: Discerning intentions in dynamic human action. Trends Cogn. Sci. 5(4), 171–178 (2001)CrossRefGoogle Scholar
- 3.Barros, J., Serra, F., Santos, V., Silva, F.: Tele-kinesthetic teaching of motion skills to humanoid robots through haptic feedback. In: IEEERAS Humanoids’ 2014 Workshop on Policy Representations for Humanoid Robots (2014)Google Scholar
- 4.Carberry, S.: Techniques for plan recognition. User Model. User-Adap. Inter. 11(1–2), 31–48 (2001)CrossRefMATHGoogle Scholar
- 5.Chater, N., Vitányi, P.: Simplicity: a unifying principle in cognitive science? Trends Cogn. Sci. 7(1), 19–22 (2003)CrossRefGoogle Scholar
- 6.Chella, A., Dindo, H., Infantino, I.: A cognitive framework for imitation learning. Robot. Auton. Syst. 54(5), 403–408 (2006)CrossRefGoogle Scholar
- 7.Dasigi, V., Reggia, J.: Parsimonious covering as a method for natural language interfaces to expert systems. AI Med. 1(1), 49–60 (1989)Google Scholar
- 8.Dindo, H., Chella, A., La Tona, G., Vitali, M., Nivel, E., Thórisson, K.R.: Learning problem solving skills from demonstration: an architectural approach. In: Schmidhuber, J., Thórisson, K.R., Looks, M. (eds.) AGI 2011. LNCS, vol. 6830, pp. 194–203. Springer, Heidelberg (2011)CrossRefGoogle Scholar
- 9.Friesen, A.L., Rao, R.P.: Imitation learning with hierarchical actions. In: 9th International Conference on Development and Learning, pp. 263–268. IEEE (2010)Google Scholar
- 10.Ghallab, M., Nau, D., Traverso, P.: Automated Planning. Elsevier, Amsterdam (2004)MATHGoogle Scholar
- 11.Haikonen, P.: The Cognitive Approach to Conscious Machines. Imprint Academic (2003)Google Scholar
- 12.Henson, C., Sheth, A., Thirunarayan, K.: Semantic perception: converting sensory observations to abstractions. Internet Comput. 16(2), 26–34 (2012). IEEECrossRefGoogle Scholar
- 13.Huang, D.-W., Katz, G.E., Langsfeld, J.D., Oh, H., Gentili, R.J., Reggia, J.A.: An object-centric paradigm for robot programming by demonstration. In: Schmorrow, D.D., Fidopiastis, C.M. (eds.) AC 2015. LNCS, vol. 9183, pp. 745–756. Springer, Heidelberg (2015)CrossRefGoogle Scholar
- 14.Jansen, B., Belpaeme, T.: A computational model of intention reading in imitation. Robot. Auton. Syst. 54(5), 394–402 (2006)CrossRefGoogle Scholar
- 15.Meltzoff, A., Moore, A.: Imitation of facial and manual gestures by human neonates. Science 198(4312), 75–78 (1977)CrossRefGoogle Scholar
- 16.Nebel, B., Koehler, J.: Plan reuse versus plan generation. Artif. Intell. 76(1), 427–454 (1995)CrossRefGoogle Scholar
- 17.Peng, Y., Reggia, J.: Abductive Inference Models for Diagnostic Problem-Solving. Springer, New York (1990)CrossRefMATHGoogle Scholar
- 18.Shivashankar, V., Alford, R., Kuter, U., Nau, D.: The GoDeL planning system: a more perfect union of domain-independent and hierarchical planning. In: Proceedings of the 23rd International Joint Conference on AI. AAAI (2013)Google Scholar
- 19.Wen, F., Chang, C.: A new approach to fault diagnosis in electrical distribution networks using a genetic algorithm. AI Eng. 12(1), 69–80 (1998)Google Scholar
- 20.Wu, Y., Su, Y., Demiris, Y.: A morphable template framework for robot learning by demonstration. Robot. Auton. Syst. 62(10), 1517–1530 (2014)CrossRefGoogle Scholar
Copyright information
© Springer International Publishing Switzerland 2016