Predictive information and explorative behavior of autonomous robots
- 1.1k Downloads
- 74 Citations
Abstract.
Measures of complexity are of immediate interest for the field of autonomous robots both as a means to classify the behavior and as an objective function for the autonomous development of robot behavior. In the present paper we consider predictive information in sensor space as a measure for the behavioral complexity of a two-wheel embodied robot moving in a rectangular arena with several obstacles. The mutual information (MI) between past and future sensor values is found empirically to have a maximum for a behavior which is both explorative and sensitive to the environment. This makes predictive information a prospective candidate as an objective function for the autonomous development of such behaviors. We derive theoretical expressions for the MI in order to obtain an explicit update rule for the gradient ascent dynamics. Interestingly, in the case of a linear or linearized model of the sensorimotor dynamics the structure of the learning rule derived depends only on the dynamical properties while the value of the MI influences only the learning rate. In this way the problem of the prohibitively large sampling times for information theoretic measures can be circumvented. This result can be generalized and may help to derive explicit learning rules from complexity theoretic measures.
PACS.
89.70.Cf Entropy and other measures of information 87.19.lo Information theory 87.85.St RoboticsReferences
- W. Bialek, I. Nemenman, N. Tishby, Neural Comput. 13, 2409 (2001) zbMATHCrossRefGoogle Scholar
- G. Box, G.M. Jenkins, G.C. Reinsel, Time Series Analysis: Forecasting and Control (Prentice Hall, 1994) Google Scholar
- T.M. Cover, J.A. Thomas, Elements of Information Theory Wiley series in telecommunications (Wiley, New York, 1991) Google Scholar
- J.P. Crutchfield, K. Young, Phys. Rev. Lett. 63, 105 (1989) CrossRefADSMathSciNetGoogle Scholar
- R. Der, Theory in Biosciences 120, 179 (2001) Google Scholar
- R. Der, F. Hesse, G. Martius, J. Adaptive Behavior 14, 105 (2005) CrossRefGoogle Scholar
- R. Der, R. Liebscher, True autonomy from self-organized adaptivity, in Proc. Workshop Biologically Inspired Robotics. The Legacy of Grey Walter 14-16 August 2002, Bristol Labs (Bristol, 2002) Google Scholar
- R. Der, G. Martius, From motor babbling to purposive actions: Emerging self-exploration in a dynamical systems approach to early robot development, in From Animals to Animats, Vol. 4095 of Lecture Notes in Computer Science, edited by S. Nolfi (Springer, 2006), p. 406 Google Scholar
- R. Der, G. Martius, F. Hesse, Let it roll – emerging sensorimotor coordination in a spherical robot, in Artificial Life X, edited by L.M. Rocha (MIT Press, 2006), p. 192 Google Scholar
- R. Der, U. Steinmetz, F. Pasemann, Homeokinesis - a new principle to back up evolution with learning, in Computational Intelligence for Modelling, Control, and Automation, Vol. 55 of Concurrent Systems Engineering Series (IOS Press, Amsterdam, 1999), p. 43 Google Scholar
- P. Grassberger, Int. J. Theor. Phys. 25(9) 907 (1986) Google Scholar
- G. Jumary, Relative Information, Vol. 47 of Springer Series in Synergetics (Springer-Verlag, Berlin Heidelberg, 1990) Google Scholar
- A.S. Klyubin, D. Polani, C.L. Nehaniv, Empowerment: A universal agent-centric measure of control, in Proc. CEC. IEEE, 2005 Google Scholar
- A.S. Klyubin, D. Polani, C.L. Nehaniv, Neural Comput. 19, 2387, 2007 Google Scholar
- M. Lungarella, G. Metta, R. Pfeifer, G. Sandini, Connect. Sci. 15(4), 151 (2003) CrossRefGoogle Scholar
- M. Lungarella, T. Pegors, D. Bulwinkle, O. Sporns, Neuroinformatics 3(3) 243 (2005) Google Scholar
- M. Lungarella, O. Sporns, Comput. Biol. 2(10), e144 (2006) Google Scholar
- G. Martius, R. Der, lpzrobots – simulation tool for autonomous robots, http://robot.informatik.uni-leipzig.de/, 2007 Google Scholar
- P.-Y. Oudeyer, F. Kaplan, V.V. Hafner, A. Whyte, The playground experiment: Task-independent development of a curious robot, in Proceedings of the AAAI Spring Symposium on Developmental Robotics, edited by D. Bank, L. Meeden (Stanford, California, 2005), p. 42 Google Scholar
- H. Risken, The Fokker-Planck Equation (Springer, 1989) Google Scholar
- J. Schmidhuber, Completely self-referential optimal reinforcement learners, in ICANN (2), pp. 223–233 (2005) Google Scholar
- R. Smith, Open dynamics engine, http://ode.org/, 2005 Google Scholar
- S. Still, Statistical mechanics approach to interactive learning arXiv:0709.1948v1 [physics.data-an], 2007. submitted Google Scholar
- A. Stout, G. Konidaris, A. Barto, Iintrinsically motivated reinforcement learning: A promising framework for developmental robotics, in The AAAI Spring Symposium on Developmental Robotics, 2005 Google Scholar
- J. Weng, J. McClelland, A. Pentland, O. Sporns, I. Stockman, M. Sur, E. Thelen, Science 291, 599 (2001) CrossRefGoogle Scholar
Copyright information
Open Access This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution and reproduction in any medium, provided the original author(s) and source are credited.