Abstract
In this paper, we review hierarchical problem networks, which encode knowledge about how to decompose planning tasks, and report an approach to learning this expertise from sample solutions. In this framework, procedural knowledge comprises a set of conditional methods that decompose problems – sets of goals – into subproblems. Problem solving involves search through a space of hierarchical plans that achieve top-level goals. Acquisition involves creation of new methods, including state conditions for when they are relevant and goal conditions for when to avoid them. We describe HPNL, a system that learns new methods by analyzing sample hierarchical plans, using violated constraints to identify state conditions and ordering conflicts to determine goal conditions. Experiments with on-line learning in three planning domains demonstrate that HPNL acquires expertise that reduces search on novel problems and examine the importance of learning goal conditions. In closing, we contrast the approach with earlier methods for acquiring search-control knowledge, including explanation-based learning and inductive logic programming. We also discuss limitations and plans for future research.
ILP 2022, 31st International Conference on Inductive Logic Programming, Cumberland Lodge, Windsor, UK.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
We constructed these constraints manually, although in principle they could have been extracted automatically from the operators’ definitions.
References
Cropper, A., Muggleton, S.H.: Learning efficient logical robot strategies involving composable objects. In: Proceedings of the Twenty-Fourth International Joint Conference on Artificial Intelligence, Buenos Aires, Argentina, pp. 3423–3429. AAAI Press (2015)
Fine-Morris, M., et al.: Learning hierarchical task networks with landmarks and numeric fluents by combining symbolic and numeric regression. In: Proceedings of the Eighth Annual Conference on Advances in Cognitive Systems. Cognitive Systems Foundation (2020)
Hogg, C., Muñoz-Avila, H., Aha, D.W.: HTN-Maker: learning HTNs with minimal additional knowledge engineering required. In: Proceedings of the Twenty-Third National Conference on Artificial Intelligence. AAAI Press (2008)
Iba, G.A.: A heuristic approach to the discovery of macro-operators. Mach. Learn. 3, 285–317 (1989)
Ilghami, O., Nau, D.S., Muñoz-Avila, H., Aha, D.W.: CaMeL: learning method preconditions for HTN planning. In: Proceedings of the Sixth International Conference on AI Planning and Scheduling, Toulouse, France, pp. 131–141. AAAI Press (2002)
Langley, P., Choi, D.: Learning recursive control programs from problem solving. J. Mach. Learn. Res. 7, 493–518 (2006)
Langley, P., Shrobe, H.E.: Hierarchical problem networks for knowledge-based planning. In: Proceedings of the Ninth Annual Conference on Advances in Cognitive Systems. Cognitive Systems Foundation (2021)
Langley, P., Simon, H.A.: Applications of machine learning and rule induction. Commun. ACM 38(November), 55–64 (1995)
Lloyd, J.W.: Foundations of logic programming. Springer, Heidelberg (1984). https://doi.org/10.1007/978-3-642-96826-6
Marsella, S.C., Schmidt, C.F.: A method for biasing the learning of nonterminal reduction rules. In: Minton, S. (ed.), Machine learning methods for planning. Morgan Kaufmann: San Francisco, CA (1993)
Minton, S.: Quantitative results concerning the utility of explanation-based learning. In: Proceedings of the Seventh National Conference on Artificial Intelligence, St. Paul, MN, pp. 564–569. Morgan Kaufmann (1988)
Nau, D., et al.: SHOP2: An HTN planning system. J. Artif. Intell. Res. 20, 379–404 (2003)
Nejati, N., Langley, P., Könik, T.: Learning hierarchical task networks by observation. In: Proceedings of the Twenty-Third International Conference on Machine Learning, Pittsburgh, PA, pp. 665–672 (2006)
Reddy, C., Tadepalli, P.: Learning goal-decomposition rules using exercises. In: Proceedings of the Fourteenth International Conference on Machine Learning, Nashville, TN, pp. 278–286. Morgan Kaufmann (1997)
Ruby, D., Kibler, D.: SteppingStone: An empirical and analytical evaluation. In: Proceedings of the Tenth National Conference on Artificial Intelligence, Menlo Park, CA, pp. 527–532. AAAI Press (1991)
Schmid, U., Kitzelmann, E.: Inductive rule learning on the knowledge level. Cogn. Syst. Res. 12, 237–248 (2011)
Shavlik, J.: Acquiring recursive and iterative concepts with explanation-based learning. Mach. Learn. 5, 39–70 (1990)
Shivashankar, V., Kuter, U., Nau, D., Alford, R.: A hierarchical goal-based formalism and algorithm for single-agent planning. In: Proceedings of the Eleventh International Conference on Autonomous Agents and Multiagent Systems, Valencia, Spain, pp. 981–988 (2012)
Sleeman, D., Langley, P., Mitchell, T.: Learning from solution paths: An approach to the credit assignment problem. AI Mag. 3, 48–52 (1982)
Acknowledgements
This research was supported by Grant N00014-20-1-2643 from the US Office of Naval Research, which is not responsible for its contents. We thank Howie Shrobe, Boris Katz, Gary Borchardt, Sue Felshin, Mohan Sridharan, and Ed Katz for discussions that influenced the ideas reported here.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2024 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Langley, P. (2024). Learning Hierarchical Problem Networks for Knowledge-Based Planning. In: Muggleton, S.H., Tamaddoni-Nezhad, A. (eds) Inductive Logic Programming. ILP 2022. Lecture Notes in Computer Science(), vol 13779. Springer, Cham. https://doi.org/10.1007/978-3-031-55630-2_6
Download citation
DOI: https://doi.org/10.1007/978-3-031-55630-2_6
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-55629-6
Online ISBN: 978-3-031-55630-2
eBook Packages: Computer ScienceComputer Science (R0)