Bilevel Planning with Learned Symbolic Abstractions from Interaction Data

2026-03-09

AI Summary

This work addresses the challenge of integrating continuous dynamic modeling with discrete symbolic reasoning in neural-symbolic systems, a gap exacerbated by the lack of effective validation for generated plans. The authors propose a two-layer neural-symbolic planning framework: at the high level, probabilistic symbolic rules learned from interaction data efficiently generate candidate plans; at the low level, a learned continuous effects model verifies plan feasibility and triggers forward search when necessary. This approach is the first to combine probabilistic symbolic abstraction with continuous feasibility validation, enabling synergistic efficient reasoning and reliable execution. Evaluated on multi-object manipulation tasks, the system substantially outperforms purely symbolic methods by accurately identifying infeasible plans, matches the overall performance of continuous forward search, and resolves most problems through efficient symbolic reasoning alone.

Technology Category

• Planning, Routing, and Scheduling: Mixed Discrete/Continuous Planning
• Cognitive Modeling & Cognitive Systems: Symbolic Representations
• Machine Learning: Neuro-Symbolic Learning

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Intelligent agents must reason over both continuous dynamics and discrete representations to generate effective plans in complex environments. Previous studies have shown that symbolic abstractions can emerge from neural effect predictors trained with a robot's unsupervised exploration. However, these methods rely on deterministic symbolic domains, lack mechanisms to verify the generated symbolic plans, and operate only at the abstract level, often failing to capture the continuous dynamics of the environment. To overcome these limitations, we propose a bilevel neuro-symbolic framework in which learned probabilistic symbolic rules generate candidate plans rapidly at the high level, and learned continuous effect models verify these plans and perform forward search when necessary at the low level. Our experiments on multi-object manipulation tasks demonstrate that the proposed bilevel method outperforms symbolic-only approaches, reliably identifying failing plans through verification, and achieves planning performance statistically comparable to continuous forward search while resolving most problems via efficient symbolic reasoning.

Problem

• bilevel planning
• symbolic abstractions
• continuous dynamics
• plan verification
• neuro-symbolic reasoning

Innovation

• bilevel planning
• neuro-symbolic
• symbolic abstraction
• probabilistic symbolic rules
• continuous effect models