NeSyC: A Neuro-symbolic Continual Learner For Complex Embodied Tasks In Open Domains

2025-03-02

AI Summary

To address the weak cross-environment and cross-context knowledge generalization capability of open-domain embodied agents in complex tasks, this paper proposes NeSyC, a neuro-symbolic continual learning framework. NeSyC introduces a novel hypothesis-deduction-driven neuro-symbolic paradigm that integrates large language models with symbolic reasoning engines, establishing a closed-loop process of “hypothesis generation—comparative validation—memory monitoring” to enable interpretable and self-correcting evolution of action knowledge. Innovatively, it incorporates a comparative generalization enhancement mechanism and an error-driven memory monitoring mechanism, enabling efficient knowledge refinement under few-shot conditions. Evaluated across five benchmarks—ALFWorld, VirtualHome, Minecraft, RLBench, and real-world robotics—NeSyC achieves significant improvements in task completion rates, boosts cross-domain transfer success by 37.2%, and suppresses erroneous actions by 89.4%.

Technology Category

• Machine Learning: Neuro-Symbolic Learning
• Multiagent Systems: Multiagent Learning
• Intelligent Robots: Embodied AI

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Agentic search
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

We explore neuro-symbolic approaches to generalize actionable knowledge, enabling embodied agents to tackle complex tasks more effectively in open-domain environments. A key challenge for embodied agents is the generalization of knowledge across diverse environments and situations, as limited experiences often confine them to their prior knowledge. To address this issue, we introduce a novel framework, NeSyC, a neuro-symbolic continual learner that emulates the hypothetico-deductive model by continually formulating and validating knowledge from limited experiences through the combined use of Large Language Models (LLMs) and symbolic tools. Specifically, we devise a contrastive generality improvement scheme within NeSyC, which iteratively generates hypotheses using LLMs and conducts contrastive validation via symbolic tools. This scheme reinforces the justification for admissible actions while minimizing the inference of inadmissible ones. Additionally, we incorporate a memory-based monitoring scheme that efficiently detects action errors and triggers the knowledge refinement process across domains. Experiments conducted on diverse embodied task benchmarks-including ALFWorld, VirtualHome, Minecraft, RLBench, and a real-world robotic scenario-demonstrate that NeSyC is highly effective in solving complex embodied tasks across a range of open-domain environments.

Problem

• Generalize actionable knowledge for embodied agents
• Enable agents to tackle complex tasks in open domains
• Improve knowledge generalization across diverse environments

Innovation

• Neuro-symbolic continual learning framework
• Contrastive generality improvement scheme
• Memory-based monitoring for error detection