Abstract
Despite Large Language Models (LLMs) have revolutionised Natural Language Processing (NLP), their capability of performing logical reasoning and automated planning is still debated. In this context, the state of the art is PlanGPT, a GPT-2 model specifically trained for planning tasks. This recent approach provides GPT-based planning policies with remarkable performance, but it can generate invalid plans containing violated action preconditions or unsatisfied goals. To address this limitation, we propose an extension of PlanGPT that integrates a plan validator into the generation process. The validator is exploited to prune invalid plan prefixes during the GPT token generation, obtaining a more robust and powerful solution to planning via GPT. We empirically evaluate the effectiveness of our approach and demonstrate its potential in various planning domains.
N. Rossetti was enrolled in the National Doctorate on AI conducted by Sapienza, University of Rome with the University of Brescia.
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
Similar content being viewed by others
References
Arora, D., Kambhampati, S.: Learning and leveraging verifiers to improve planning capabilities of pre-trained language models. CoRR abs/2305.17077 (2023)
Chiari, M., Gerevini, A.E., Loreggia, A., Putelli, L., Serina, I.: Fast and slow goal recognition. In: Dastani, M., Sichman, J.S., Alechina, N., Dignum, V. (eds.) Proceedings of the 23rd International Conference on Autonomous Agents and Multiagent Systems, AAMAS 2024, Auckland, New Zealand, 6–10 May 2024, pp. 354–362. ACM (2024)
Chiari, M., Gerevini, A.E., Percassi, F., Putelli, L., Serina, I., Olivato, M.: Goal recognition as a deep learning task: the GRNet approach. In: Koenig, S., Stern, R., Vallati, M. (eds.) Proceedings of the Thirty-Third International Conference on Automated Planning and Scheduling, Prague, Czech Republic, 8–13 July 2023, pp. 560–568. AAAI Press (2023)
Devlin, J., Chang, M., Lee, K., Toutanova, K.: BERT: pre-training of deep bidirectional transformers for language understanding. In: NAACL-HLT (1), pp. 4171–4186. Association for Computational Linguistics (2019)
Gerevini, A., Saetti, A., Serina, I.: Planning through stochastic local search and temporal action graphs in LPG. J. Artif. Intell. Res. 20, 239–290 (2003). https://doi.org/10.1613/jair.1183
Gerevini, A., Serina, I.: LPG: a planner based on local search for planning graphs with action costs. In: AIPS, pp. 13–22. AAAI Press (2002)
Geva, M., Khashabi, D., Segal, E., Khot, T., Roth, D., Berant, J.: Did Aristotle use a laptop? A question answering benchmark with implicit reasoning strategies. Trans. Assoc. Comput. Linguistics 9, 346–361 (2021)
Howey, R., Long, D., Fox, M.: VAL: automatic plan validation, continuous effects and mixed initiative planning using PDDL. In: ICTAI, pp. 294–301. IEEE Computer Society (2004)
Hu, Y., De Giacomo, G.: Generalized planning: synthesizing plans that work for multiple environments. In: IJCAI, pp. 918–923. IJCAI Org (2011)
Jiang, Z., Xu, F.F., Araki, J., Neubig, G.: How can we know what language models know. Trans. Assoc. Comput. Linguistics 8, 423–438 (2020)
McDermott, D., G., et al.: PDDL-the planning domain definition language (1998). https://api.semanticscholar.org/CorpusID:59656859
Pallagani, V., et al.: Plansformer tool: demonstrating generation of symbolic plans using transformers. In: IJCAI, pp. 7158–7162. IJCAI Org. (2023)
Petroni, F., et al.: Language models as knowledge bases? In: EMNLP/IJCNLP (1), pp. 2463–2473. Association for Computational Linguistics (2019)
Radford, A., Narasimhan, K.: Improving language understanding by generative pre-training. In: preprint (2018). api.semanticscholar.org/CorpusID:49313245
Rossetti, N., et al.: Learning general policies for planning through GPT models. In: Proceedings of the International Conference on Automated Planning and Scheduling, vol. 34, no. 1, pp. 500–508 (2024)
Seipp, J., Torralba, A., Hoffmann, J.: PDDL generators (2022). https://github.com/AI-Planning/pddl-generators
Serina, L., Chiari, M., Gerevini, A.E., Putelli, L., Serina, I.: A preliminary study on BERT applied to automated planning. In: IPS/AI*IA, vol. 3345. CEUR-WS.org (2022)
Srivastava, S., Immerman, N., Zilberstein, S.: Learning generalized plans using abstract counting. In: AAAI, pp. 991–997. AAAI Press (2008)
Srivastava, S., Immerman, N., Zilberstein, S.: A new representation and associated algorithms for generalized planning. Artif. Intell. 175(2), 615–647 (2011)
Valmeekam, K., Hernandez, A.O., Sreedharan, S., Kambhampati, S.: Large language models still can’t plan (A benchmark for LLMs on planning and reasoning about change). CoRR abs/2206.10498 (2022)
Valmeekam, K., Sreedharan, S., Marquez, M., Hernandez, A.O., Kambhampati, S.: On the planning abilities of large language models (A critical investigation with a proposed benchmark). CoRR abs/2302.06706 (2023)
Vaswani, A., S., et al.: Attention is all you need. In: NIPS, pp. 5998–6008. Curran Associates Inc. (2017)
Wang, Y., Wang, W., Joty, S.R., Hoi, S.C.H.: CodeT5: identifier-aware unified pre-trained encoder-decoder models for code understanding and generation. In: EMNLP (1), pp. 8696–8708. Association for Computational Linguistics (2021)
Acknowledgements
This work was supported by EU H2020 project AIPlan4EU (GA 101016442), EU ICT-48 2020 project TAILOR (GA 952215), MUR PRIN project RIPER (No. 20203FFYLK), Climate Change AI project (No. IG-2023-174), and Regione Lombardia through the initiative “Il Piano Lombardia - Interventi per la ripresa economica”.
Author information
Authors and Affiliations
Corresponding authors
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
Rossetti, N., Tummolo, M., Gerevini, A.E., Olivato, M., Putelli, L., Serina, I. (2024). Enhancing GPT-Based Planning Policies by Model-Based Plan Validation. In: Besold, T.R., d’Avila Garcez, A., Jimenez-Ruiz, E., Confalonieri, R., Madhyastha, P., Wagner, B. (eds) Neural-Symbolic Learning and Reasoning. NeSy 2024. Lecture Notes in Computer Science(), vol 14980. Springer, Cham. https://doi.org/10.1007/978-3-031-71170-1_26
Download citation
DOI: https://doi.org/10.1007/978-3-031-71170-1_26
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-71169-5
Online ISBN: 978-3-031-71170-1
eBook Packages: Computer ScienceComputer Science (R0)