Abstract
Intrinsic motivations can be defined by contrasting them to extrinsic motivations. Extrinsic motivations are directed to drive the learning of behavior directed to satisfy basic needs related to the organisms’ survival and reproduction. Intrinsic motivations, instead, are motivations that serve the evolutionary function of acquiring knowledge (e.g., the capacity to predict) and competence (i.e., the capacity to do) in the absence of extrinsic motivations: this knowledge and competence can be later exploited for producing behaviors that enhance biological fitness. Knowledge-based intrinsic motivation mechanisms (KB-IM), usable for guiding learning on the basis of the level or change of knowledge, have been widely modeled and studied. Instead, competence-based intrinsic motivation mechanisms (CB-IM), usable for guiding learning on the basis of the level or improvement of competence, have been much less investigated. The goal of this chapter is twofold. First, it aims to clarify the nature and possible roles of CB-IM mechanisms for learning, in particular in relation to the cumulative acquisition of a repertoire of skills. Second, it aims to review a specific CB-IM mechanism, the Temporal-Difference Competence-Based Intrinsic Motivation (TD-CB-IM). TD-CB-IM measures the improvement rate of skill acquisition on the basis of the Temporal-Difference learning signal (TD error) that is used in several reinforcement learning (RL) models. The effectiveness of the mechanism is supported by reviewing and discussing in depth the results of experiments in which the TD-CB-IM mechanism is successfully exploited by a hierarchical RL model controlling a simulated navigating robot to decide when to train different skills in different environmental conditions.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
Reinforcement learning models mimic the trial-and-error learning processes of animals directed to achieve an extrinsic reward, in particular those studied by behaviorist psychology with instrumental learning paradigms (Lieberman 1993) (but the models are also used to capture some mechanisms of Pavlovian learning). One of the most biologically plausible RL models, the actor-critic RL model (Houk et al. 1995; Sutton and Barto 1998), is formed by (a) an actor, which progressively learns to select actions so to maximize rewards, and (b) a critic, which progressively learns to assign an evaluation (an estimate of future rewards) to each state on the basis of the received rewards (the actor is trained to act so as to move the agent toward states with higher evaluations).
- 2.
By “hierarchical” we mean that some components of the system, usually processing information at a more abstract level, exert an influence on other components, usually processing information at more detailed level. By “modular” we mean that different chunks of behavior are encoded in different portions of the system. Modularity can be either “structural,” that is, related to strong connections within groups of neurons and looser connections between groups, or “functional,” for example leading to encode different chunks of behavior within different portions of a rather homogeneous system on the basis of specific self-organizing mechanisms.
References
Baldassarre, G.: A modular neural-network model of the basal ganglia’s role in learning and selecting motor behaviours. In: Altmann, E.M., Cleermans, A., Schunn, C.D, Gray, W.D. (eds.) Proceedings of the Fourth International Conference on Cognitive Modeling (ICCM2001), pp. 37–42. Fairfax, Virgina, USA, 26–29 July 2001. Lawrence Erlbaum, Mahwah (2001)
Baldassarre, G.: A modular neural-network model of the basal ganglia’s role in learning and selecting motor behaviours. J. Cogn. Syst. Res. 3(2), 5–13. Special Issue Dynamic and Recurrent Neural Networks (2002a)
Baldassarre, G.: Planning with neural networks and reinforcement learning. Ph.D. Thesis, Computer Science Department, University of Essex, Colchester, UK (2002b)
Baldassarre, G.: What are intrinsic motivations? a biological perspective. In: Cangelosi, A., Triesch, J., Fasel, I., Rohlfing, K., Nori, F., Oudeyer, P.-Y., Schlesinger, M, Nagai, Y. (eds.) Proceedings of the International Conference on Development and Learning and Epigenetic Robotics (ICDL-EpiRob-2011), pp. E1–E8. Frankfurt, Germany, 24–27 August, 2011. IEEE, Piscataway (2011)
Baldassarre, G., Mannella, F., Fiore, V.G., Redgrave, P., Gurney, K., Mirolli, M.: Intrinsically motivated action-outcome learning and goal-based action recall: A system-level bio-constrained computational model. Neural Netw. (2012, in press)
Baldassarre, G., Mirolli, M.: What are the key open challenges for understanding the autonomous cumulative learning of skills? The Newslett. Auton. Mental Dev. Techn. Comm. (IEEE CIS AMD Newslett.) 7(1), 11 (2010)
Barto, A., Singh, S., Chentanez, N.: Intrinsically motivated learning of hierarchical collections of skills. In: International Conference on Developmental Learning (ICDL2004). La Jolla, CA, 20–22 October, 2004. IEEE, Piscataway (2004)
Barto, A.G., Mahadevan, S.: Recent advances in hierarchical reinforcement learning. Discr. Event Dyn. Syst. 13(4), 341–379 (2003)
Botvinick, M.M., Niv, Y., Barto, A.: Hierarchically organized behavior and its neural foundations: A reinforcement-learning perspective. Cognition 113(3), 262–280 (2008)
Caligiore, D., Mirolli, M., Parisi, D., Baldassarre, G.: A bio-inspired hierarchical reinforcement learning architecture for modeling learning of multiple skills with continuous state and actions. In: Kuipers, B., Shultz, T., Stoytchev, A., Yu, C. (eds.) IEEE International Conference on Development and Learning (ICDL2010). Ann Arbor, MI, USA, 18–21 August, 2010 IEEE, Piscataway (2010)
Deci, E., Koestner, R., Ryan, R.: Extrinsic rewards and intrinsic motivation in education: Reconsidered once again. Rev. Educ. Res. 71(1), 1–27 (2001)
Doya, K., Samejima, K., Katagiri, K.-I, Kawato, M.: Multiple model-based reinforcement learning. Neural Comput. 14(6), 1347–1369 (2002)
Elfwing, S., Uchibe, E., Doya, K., Christensen, H.: Evolutionary development of hierarchical learning structures. IEEE Trans. Evol. Comput. 11(2), 249–264 (2007)
Harlow, H.F.: Learning and satiation of response in intrinsically motivated complex puzzle performance by monkeys. J. Comp. Physiol. Psychol. 43, 289–294 (1950)
Hart, S., Grupen, R.: Learning generalizable control programs. IEEE Trans. Auton. Mental Dev. 3(1), 216–231 (2011)
Hart, S., Grupen, R.: Intrinsically motivated affordance discovery and modeling. In: Baldassarre, G., Mirolli, M. (eds.) Intrinsically Motivated Learning in Natural and Artificial Systems. Springer, Berlin (2012, this volume)
Houk, J.C., Adams, J.L., Barto, A.G.: A model of how the basal ganglia generate and use neural signals that predict reinforcement. In: Houk, J.C., Davids, J.L., Beiser, D.G. (eds.) Models of Information Processing in the Basal Ganglia, pp. 249–270. The MIT Press, Cambridge (1995)
Jacobs, R., Jordan, M., Nowlan, S., Hinton, G.: Adaptive mixtures of local experts. Neural Comput. 3(1), 79–87 (1991)
Jung, T., Polani, D., Stone, P.: Empowerment for continuous agent-environment systems. Adap. Behav. 19(1), 16–39 (2011)
Kakade, S., Dayan, P.: Dopamine: Generalization and bonuses. Neural Netw. 15(4–6), 549–559 (2002)
Kaplan, F., Oudeyer, P.-Y.: In: Search of the neural circuits of intrinsic motivation. Front. Neurosci. 1, 225–236 (2007)
Klyubin, A., Polani, D., Nehaniv, C.: Empowerment: A universal agent-centric measure of control. In: The 2005 IEEE Congress on Evolutionary Computation, vol. 1, pp. 128–135. Edinburg UK, 2–4 September, (2005)
Lieberman, D.A.: Learning, Behaviour and Cognition. Pacific Grove, CA: Brooks/Cole (1993)
Luciw, M., Graziano, V., Ring, M., Schmidhuber, J.: Artificial curiosity with planning for autonomous perceptual and cognitive development. In: Cangelosi, A., Triesch, J., Fasel, I., Rohlfing, K., Nori, F., Oudeyer, P.-Y., Schlesinger, M., Nagai, Y. (eds.) IEEE International Conference on Development and Learning (ICDL2011), pp. E1–8. IEEE, Frankfurt, Germany, 24–27 August, 2011. Piscataway (2011)
McCloskey, M., Cohen, N.: Catastrophic interference in connectionist networks: The sequential learning problem. In: Bower, G.H. (ed.) The Psychology of Learning and Motivation, vol. 24, pp. 109–165. Academic Press, San Diego (1989)
McGovern, A., Barto, A.: Automatic discovery of subgoals in reinforcement learning using diverse density. Technical report of the faculty publication series, University of Massachusetts – Amherst, Computer Science Department (2001)
Meunier, D., Lambiotte, R., Bullmore, E.T.: Modular and hierarchically modular organization of brain networks. Front. Neurosci. 4, 200 (2010)
Mirolli, M., Baldassarre, G.: Functions and mechanisms of intrinsic motivations: The knowledge versus competence distinction. In: Baldassarre, G., Mirolli, M. (eds.) Intrinsically Motivated Learning in Natural and Artificial Systems. Springer, Berlin (2012, this volume)
Mirolli, M., Santucci, V.G., Baldassarre, G.: Phasic dopamine as a prediction error of intrinsic and extrinsic reinforcement driving both action acquisition and reward maximization: A simulated robotic study. Neural Netw. (2012, submitted)
Oudeyer, P.-Y., Banares, A., Frédéric, K.: Intrinsically motivated learning of real world sensorimotor skills with developmental constraints. In: Baldassarre, G., Mirolli, M. (eds.) Intrinsically Motivated Learning in Natural and Artificial Systems. Springer, Berlin (2012, this volume)
Oudeyer, P.-Y., Kaplan, F.: What is intrinsic motivation? a typology of computational approaches. Front. Neurorobot. 1, 6 (2007)
Oudeyer, P.-Y., Kaplan, F., Hafner, V.V.: Intrinsic motivation systems for autonomous mental development. IEEE Trans. Evol. Comput. 11(2), 265–286 (2007)
Pickett, M., Barto, A.: Policyblocks: An algorithm for creating useful macro-actions in reinforcement learning. In: Sammut, C., Hoffmann, A.G. (eds.) Proceedings of the Nineteenth International Conference on Machine Learning, pp. 506–513. Sydney, Australia, 8–12 July 2002. Morgan Kaufmann, San Francisco (2002)
Redgrave, P., Gurney, K.: The short-latency dopamine signal: A role in discovering novel actions? Nat. Rev. Neurosci. 7(12), 967–975 (2006)
Redgrave, P., Gurney, K., Stafford, T., Thirkettle, M., Lewis, J.: The role of the basal ganglia in discovering novel actions. In: Baldassarre, G., Mirolli, M. (eds.) Intrinsically Motivated Learning in Natural and Artificial Systems. Springer, Berlin (2012, this volume)
Ryan, R., Deci, E.: Intrinsic and extrinsic motivations: Classic definitions and new directions. Contemp. Educ. Psychol. 25, 54–67 (2000)
Santucci, V.G., Baldassarre, G., Mirolli, M.: Biological cumulative learning through intrinsic motivations: A simulated robotic study on development of visually-guided reaching. In: Johansson, B., Sahin, E., Balkenius, C. (eds.) Proceedings of the Tenth International Conference on Epigenetic Robotics (EpiRob2010), pp. 121–128. Lund, Sweden. Lund: Lund University Cognitive Studies vol.149 (2010)
Schembri, M., Mirolli, M., Baldassarre, G.: Evolution and learning in an intrinsically motivated reinforcement learning robot. In: Almeida e Costa Fernando, Rocha, L.M., Costa, E., Harvey, I., Coutinho, A. (eds.) Advances in Artificial Life. Proceedings of the 9th European Conference on Artificial Life (ECAL2007), Lisbon, Portugal, 10–14 September 2007. Lecture Notes in Artificial Intelligence, vol. 4648, pp. 294–333. Springer, Berlin (2007a)
Schembri, M., Mirolli, M., Baldassarre, G.: Evolving childhood’s length and learning parameters in an intrinsically motivated reinforcement learning robot. In: Berthouze, L., Dhristiopher, P.G., Littman, M., Kozima, H., Balkenius, C. (eds.) Proceedings of the Seventh International Conference on Epigenetic Robotics, vol. 134, pp. 141–148. Lund, Sweden. Lund: Lund University Cognitive Studies vol. 149 (2007b)
Schembri, M., Mirolli, M., Baldassarre, G.: Evolving internal reinforcers for an intrinsically motivated reinforcement-learning robot. In: Demiris, Y., Mareschal, D., Scassellati, B., Weng, J. (eds.) Proceedings of the 6th International Conference on Development and Learning, pp. E1–6. London, UK, 11–13 July 2007. IEEE, Piscataway (2007c)
Schmidhuber, J.: Curious model-building control systems. In: Proceedings of the International Joint Conference on Neural Networks, vol. 2, pp. 1458–1463 Singapore, 18–21 November (1991a)
Schmidhuber, J.: A possibility for implementing curiosity and boredom in model-building neural controllers. In: Meyer, J.-A., Wilson, S. (eds.) From Animals to Animats: Proceedings of the First International Conference on Simulation of Adaptive Behavior, Paris, France, December, 1990 pp. 222–227, MIT, Cambridge (1991b)
Schmidhuber, J.: Formal theory of creativity, fun, and intrinsic motivation (1990–2010): IEEE Trans. Auton. Mental Dev. 2(3), 230–247 (2010)
Schmidhuber, J.: Maximizing fun by creating data with easily reducible subjective complexity. In: Baldassarre, G., Mirolli, M. (eds.) Intrinsically Motivated Learning in Natural and Artificial Systems. Springer, Berlin (2012, this volume)
Schultz, W.: Getting formal with dopamine and reward. Neuron 36(2), 241–263 (2002)
Singh, S., Barto, A., Chentanez, N.: Intrinsically motivated reinforcement learning. In: Saul, L.K., Weiss, Y., Bottou, L. (eds.). Advances in Neural Information Processing Systems 17: Proceedings of the 2004 Conference. Vancouver, British Columbia, Canada, 13–18 December 2004. MIT, Cambridge (2005)
Singh, S., Lewis, R., Barto, A., Sorg, J.: Intrinsically motivated reinforcement learning: An evolutionary perspective. IEEE Trans. Auton. Mental Dev. 2(2), 70–82 (2010)
Stout, A., Barto, A.G.: Competence progress intrinsic motivation. In: Kuipers, B., Shultz, T., Stoytchev, A., Yu, C. (eds.) IEEE International Conference on Development and Learning (ICDL2010). Ann Arbor, MI, USA, 18–21 August, 2010. IEEE, Piscataway (2010)
Sutton, R., Precup, D., Singh, S.: Between MDPs and semi-MDPs: A framework for temporal abstraction in reinforcement learning. Artif. Intell. 112, 181–211 (1999)
Sutton, R.S., Barto, A.G.: Reinforcement Learning: An Introduction. MIT, Cambridge (1998)
Taylor, M., Stone, P.: Transfer learning for reinforcement learning domains: A survey. J. Mach. Learn. Res. 10, 1633–1685 (2009)
Thrun, S., Schwartz, A.: Finding structure in reinforcement learning. In: Tesauro, G., Touretzky, D, Leen, T. (eds.) Advances in Neural Information Processing Systems 7 (NIPS1994), Denver, Colorado, USA, pp. 385–392. MIT, Cambridge (1995)
Vigorito, C., Barto, A.: Intrinsically motivated hierarchical skill learning in structured environments. IEEE Trans. Auton. Mental Dev. 2(2), 132–143 (2010)
von Hofsten, C.: Action in development. Dev. Sci. 10(1), 54–60 (2007)
Vygotsky, L.S.: Mind in society: The development of higher psychological processes. Cambridge, MA: Harvard University Press (1978)
White, R.W.: Motivation reconsidered: The concept of competence. Psychol. Rev. 66, 297–333 (1959)
Yao, X.: Evolving artificial neural networks. In: Proceedings of the IEEE, vol. 87, pp. 1423–1447. (1999)
Acknowledgements
This research has received funds from the European Commission 7th Framework Programme (FP7/2007-2013), “Challenge 2 - Cognitive Systems, Interaction, Robotics,” Grant Agreement No. ICT-IP-231722, Project “IM-CLeVeR - Intrinsically Motivated Cumulative Learning Versatile Robots.”
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2013 Springer-Verlag Berlin Heidelberg
About this chapter
Cite this chapter
Baldassarre, G., Mirolli, M. (2013). Deciding Which Skill to Learn When: Temporal-Difference Competence-Based Intrinsic Motivation (TD-CB-IM). In: Baldassarre, G., Mirolli, M. (eds) Intrinsically Motivated Learning in Natural and Artificial Systems. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-32375-1_11
Download citation
DOI: https://doi.org/10.1007/978-3-642-32375-1_11
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-32374-4
Online ISBN: 978-3-642-32375-1
eBook Packages: Computer ScienceComputer Science (R0)