Act-to-Ground: A Framework for Symbol Grounding in Planning Domains

Panagiotis Lymperopoulos, Liping Liu
Proceedings of The 19th International Conference on Neurosymbolic Learning and Reasoning, PMLR 284:777-795, 2025.

Abstract

Neurosymbolic decision-making agents inherit many of the critical transparency and interpretability benefits of planning-based symbolic agents but also face one of their central challenges: the Symbol Grounding Problem (SGP). Grounding hand-crafted symbolic planning domains to percepts typically requires training models with extensive annotated data which hinders their applicability to broader problems. In this work we propose Act-to-Ground (A2G), a framework for training grounding models for symbolic planners with weak supervision obtained through environment interaction or demonstrations. We first cast the grounding problem as an inference problem and 1) use satisfiability-based planning to provide weak supervision to the grounding model by exploiting knowledge already built into the planning domain, 2) propose an MCMC sampler that enables sampling weak labels for grounding planners, 3) improve neurosymbolic grounding performance via a score-matching objective and 4) propose a learnability condition for learning grounding models for planners.

Cite this Paper

BibTeX
@InProceedings{pmlr-v284-lymperopoulos25a, title = {Act-to-Ground: A Framework for Symbol Grounding in Planning Domains}, author = {Lymperopoulos, Panagiotis and Liu, Liping}, booktitle = {Proceedings of The 19th International Conference on Neurosymbolic Learning and Reasoning}, pages = {777--795}, year = {2025}, editor = {H. Gilpin, Leilani and Giunchiglia, Eleonora and Hitzler, Pascal and van Krieken, Emile}, volume = {284}, series = {Proceedings of Machine Learning Research}, month = {08--10 Sep}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v284/main/assets/lymperopoulos25a/lymperopoulos25a.pdf}, url = {https://proceedings.mlr.press/v284/lymperopoulos25a.html}, abstract = {Neurosymbolic decision-making agents inherit many of the critical transparency and interpretability benefits of planning-based symbolic agents but also face one of their central challenges: the Symbol Grounding Problem (SGP). Grounding hand-crafted symbolic planning domains to percepts typically requires training models with extensive annotated data which hinders their applicability to broader problems. In this work we propose Act-to-Ground (A2G), a framework for training grounding models for symbolic planners with weak supervision obtained through environment interaction or demonstrations. We first cast the grounding problem as an inference problem and 1) use satisfiability-based planning to provide weak supervision to the grounding model by exploiting knowledge already built into the planning domain, 2) propose an MCMC sampler that enables sampling weak labels for grounding planners, 3) improve neurosymbolic grounding performance via a score-matching objective and 4) propose a learnability condition for learning grounding models for planners.} }
Endnote
%0 Conference Paper %T Act-to-Ground: A Framework for Symbol Grounding in Planning Domains %A Panagiotis Lymperopoulos %A Liping Liu %B Proceedings of The 19th International Conference on Neurosymbolic Learning and Reasoning %C Proceedings of Machine Learning Research %D 2025 %E Leilani H. Gilpin %E Eleonora Giunchiglia %E Pascal Hitzler %E Emile van Krieken %F pmlr-v284-lymperopoulos25a %I PMLR %P 777--795 %U https://proceedings.mlr.press/v284/lymperopoulos25a.html %V 284 %X Neurosymbolic decision-making agents inherit many of the critical transparency and interpretability benefits of planning-based symbolic agents but also face one of their central challenges: the Symbol Grounding Problem (SGP). Grounding hand-crafted symbolic planning domains to percepts typically requires training models with extensive annotated data which hinders their applicability to broader problems. In this work we propose Act-to-Ground (A2G), a framework for training grounding models for symbolic planners with weak supervision obtained through environment interaction or demonstrations. We first cast the grounding problem as an inference problem and 1) use satisfiability-based planning to provide weak supervision to the grounding model by exploiting knowledge already built into the planning domain, 2) propose an MCMC sampler that enables sampling weak labels for grounding planners, 3) improve neurosymbolic grounding performance via a score-matching objective and 4) propose a learnability condition for learning grounding models for planners.
APA
Lymperopoulos, P. & Liu, L.. (2025). Act-to-Ground: A Framework for Symbol Grounding in Planning Domains. Proceedings of The 19th International Conference on Neurosymbolic Learning and Reasoning, in Proceedings of Machine Learning Research 284:777-795 Available from https://proceedings.mlr.press/v284/lymperopoulos25a.html.

Related Material