Abstract
Deep Learning (DL) allowed the field of Multi-Agent Reinforcement Learning (MARL) to make significant advances, speeding-up the progress in the field. However, agents trained by means of DL in MARL settings have an important drawback: their policies are extremely hard to interpret, not only at the individual agent level, but also (and especially) considering the fact that one has to take into account the interactions across the whole set of agents. In this work, we make a step towards achieving interpretability in MARL tasks. To do that, we present an approach that combines evolutionary computation (i.e., grammatical evolution) and reinforcement learning (Q-learning), which allows us to produce agents that are, at least to some extent, understandable. Moreover, differently from the typically centralized DL-based approaches (and because of the possibility to use a replay buffer), in our method we can easily employ Independent Q-learning to train a team of agents, which facilitates robustness and scalability. By evaluating our approach on the Battlefield task from the MAgent implementation in the PettingZoo library, we observe that the evolved team of agents is able to coordinate its actions in a distributed fashion, solving the task in an effective way.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
In the rest of this paper, we will define as an interpretable system one that can be understood and inspected by humans [2].
- 2.
https://www.pettingzoo.ml/magent/battlefield (accessed on 02/02/2022).
References
OroojlooyJadid, A., Hajinezhad, D.: A review of cooperative multi-agent deep reinforcement learning (2020) . arXiv:1908.03963
Barredo Arrieta, A., et al.: Explainable Artificial Intelligence (XAI): Concepts, taxonomies, opportunities and challenges toward responsible AI. Inf. Fus. 58, 82–115 (2020)
Rudin, C.: Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead. Nat. Mach. Intell. 1(5), 206–215 (2019)
Rudin, C., Radin, J.: Why are we using black box models in AI when we don’t need To? A lesson from an explainable ai competition. Harvard Data Sci. Rev .1(2) (November 2019)
Rudin, C., Chen, C., Chen, Z., Huang, H., Semenova, L., Zhong, C.: Interpretable machine learning: fundamental principles and 10 grand challenges, July 2021. arXiv:2103.11251
Custode, L.L., Iacca, G.: Evolutionary learning of interpretable decision trees (2020)
Potter, M.A., De Jong, K.A.: A cooperative coevolutionary approach to function optimization. In: Davidor, Y., Schwefel, H.-P., Männer, R. (eds.) PPSN 1994. LNCS, vol. 866, pp. 249–257. Springer, Heidelberg (1994). https://doi.org/10.1007/3-540-58484-6_269
Zheng, L., et al.: MAgent: a many-agent reinforcement learning platform for artificial collective intelligence. In: Proceedings of the AAAI Conference on Artificial Intelligence, vol. 32, pp. 8222–8223 (2018)
Terry, J.K., et al.: Pettingzoo: gym for multi-agent reinforcement learning (2020). arXiv:2009.14471
Busoniu, L., Babuska, R., De Schutter, B.: A comprehensive survey of multiagent reinforcement learning. IEEE Trans. Syst. Man Cybernet. Part C (Applications and Reviews) 38(2) 156–172 (2008)
Stone, P., Veloso, M.: Multiagent Systems: A Survey from a Machine Learning Perspective: Technical report. Defense Technical Information Center, Fort Belvoir, VA, December 1997
Yu, C., Liu, J., Nemati, S.: Reinforcement Learning in Healthcare: a survey, April 2020. arXiv:1908.08796
Sandholm, T.W., Crites, R.H.: On multiagent Q-learning in a semi-competitive domain. In: Weiß, G., Sen, S. (eds.) IJCAI 1995. LNCS, vol. 1042, pp. 191–205. Springer, Heidelberg (1996). https://doi.org/10.1007/3-540-60923-7_28
Littman, M.L.: Markov games as a framework for multi-agent reinforcement learning. In: Machine Learning Proceedings 1994. Morgan Kaufmann, San Francisco (CA), pp. 157–163 (1994)
Haynes, T., Wainwright, R.L., Sen, S., Schoenefeld, D.A.: Strongly typed genetic programming in evolving cooperation strategies. In: International Conference on Genetic Algorithms, San Francisco, CA, USA, pp. 271–278. Morgan Kaufmann Publishers Inc. (July 1995)
Tan, M.: In: Multi-agent Reinforcement Learning: Independent vs, pp. 487–494. Cooperative Agents. Morgan Kaufmann Publishers Inc., San Francisco (1997)
Lauer, M., Riedmiller, M.A.: An algorithm for distributed reinforcement learning in cooperative multi-agent systems. In: International Conference on Machine Learning, San Francisco, CA, USA, pp. 535–542. Morgan Kaufmann Publishers Inc. (2000)
Fuji, T., Ito, K., Matsumoto, K., Yano, K.: Deep multi-agent reinforcement learning using DNN-weight evolution to optimize supply chain performance. In: Hawaii International Conference on System Sciences, pp. 1278–1287. Honolulu, HI, USA, HICSS, (2018)
Omidshafiei, S., Pazis, J., Amato, C., How, J.P., Vian, J.: Deep decentralized multi-task multi-agent reinforcement learning under partial observability. In: International Conference on Machine Learning, pp. 2681–2690. Sydney, NSW, Australia, JMLR.org, August 2017
Matignon, L., Laurent, G.J., Le Fort-Piat, N.: Hysteretic q-learning: an algorithm for decentralized reinforcement learning in cooperative multi-agent teams. In: International Conference on Intelligent Robots and Systems, pp. 64–69. New York, NY, USA, IEEE/RSJ (2007)
Tampuu, A., et al.: Multiagent cooperation and competition with deep reinforcement learning, November 2015. arXiv:1511.08779
Chu, X., Ye, H.: Parameter sharing deep deterministic policy gradient for cooperative multi-agent reinforcement learning, October 2017. arXiv:1710.00336
Singh, A., Jain, T., Sukhbaatar, S.: Learning when to communicate at scale in multiagent cooperative and competitive tasks (2018). arXiv:1812.09755
Macua, S.V., et al.: Diff-DAC: distributed actor-critic for average multitask deep reinforcement learning (2019). arXiv:1710.10363
Sunehag, P., et al.: Value-decomposition networks for cooperativae multi-agent learning based on team reward. In: International Conference on Autonomous Agents and MultiAgent Systems, Stockholm, Sweden, International Foundation for Autonomous Agents and Multiagent Systems, pp. 2085–2087, July 2018
Yang, J., Nakhaei, A., Isele, D., Fujimura, K., Zha, H.: CM3: cooperative multi-goal multi-stage multi-agent reinforcement learning, January 2020. arXiv:1809.05188
Virgolin, M., De Lorenzo, A., Medvet, E., Randone, F.: Learning a formula of interpretability to learn interpretable formulas. In: Bäck, T., et al. (eds.) Parallel Problem Solving from Nature, pp. 79–93. Springer International Publishing, Cham (2020)
Barceló, P., Monet, M., Pérez, J., Subercaseaux, B.: Model interpretability through the lens of computational complexity. In: Proceedings of 33rd conference on Advances in Neural Information Processing Systems (2020)
Custode, L.L., Iacca, G.: A co-evolutionary approach to interpretable reinforcement learning in environments with continuous action spaces. In: 2021 IEEE Symposium Series on Computational Intelligence (SSCI), pp. 1–8, December 2021
Ryan, C., Collins, J.J., Neill, M.O.: Grammatical evolution: evolving programs for an arbitrary language. In: Banzhaf, W., Poli, R., Schoenauer, M., Fogarty, T.C. (eds.) EuroGP 1998. LNCS, vol. 1391, pp. 83–96. Springer, Heidelberg (1998). https://doi.org/10.1007/BFb0055930
Sutton, R.S., Barto, A.G.: Reinforcement Learning: An Introduction. A Bradford Book, Cambridge (2018)
Foerster, J., Assael, I.A., de Freitas, N., Whiteson, S.: Learning to communicate with deep multi-agent reinforcement learning. In Lee, D., Sugiyama, M., Luxburg, U., Guyon, I., Garnett, R., eds.: Advances in Neural Information Processing Systems, vol. 29, Curran Associates, Inc. Red Hook (2016)
Lotito, Q.F., Custode, L.L., Iacca, G.: A signal-centric perspective on the evolution of symbolic communication. In: Proceedings of the Genetic and Evolutionary Computation Conference. Association for Computing Machinery, pp. 120–128. New York, NY, USA, June (2021)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2022 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Crespi, M., Custode, L.L., Iacca, G. (2022). Towards Interpretable Policies in Multi-agent Reinforcement Learning Tasks. In: Mernik, M., Eftimov, T., Črepinšek, M. (eds) Bioinspired Optimization Methods and Their Applications. BIOMA 2022. Lecture Notes in Computer Science, vol 13627. Springer, Cham. https://doi.org/10.1007/978-3-031-21094-5_19
Download citation
DOI: https://doi.org/10.1007/978-3-031-21094-5_19
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-21093-8
Online ISBN: 978-3-031-21094-5
eBook Packages: Computer ScienceComputer Science (R0)