Few-Shot Neuro-Symbolic Imitation Learning for Long-Horizon Planning and Acting

2025-08-29

AI Summary

To address the low data efficiency and poor generalization of imitation learning in long-horizon tasks, this paper proposes a neuro-symbolic imitation learning framework. The method models high-level task structure as a graph, automatically mines symbolic rules via Answer Set Programming, and employs a higher-order oracle to select task-relevant variables—enhancing interpretability and cross-task transferability. At the lower level, it learns continuous control policies using diffusion-based policy learning, integrated with graph neural networks and neuro-symbolic collaborative reasoning to accurately capture non-spatial and temporal dependencies. Evaluated on six robotic control tasks, the framework achieves efficient learning from only five demonstrations, significantly improving zero-shot and few-shot generalization performance while ensuring decision transparency and traceability.

Technology Category

• Machine Learning: Neuro-Symbolic Learning
• Intelligent Robots: Learning & Optimization for ROB
• Computer Vision: Visual Reasoning & Symbolic Representations

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Economics, Online Markets and Human Computation: Humans versus LLMs for data annotation and labeling

Abstract

Imitation learning enables intelligent systems to acquire complex behaviors with minimal supervision. However, existing methods often focus on short-horizon skills, require large datasets, and struggle to solve long-horizon tasks or generalize across task variations and distribution shifts. We propose a novel neuro-symbolic framework that jointly learns continuous control policies and symbolic domain abstractions from a few skill demonstrations. Our method abstracts high-level task structures into a graph, discovers symbolic rules via an Answer Set Programming solver, and trains low-level controllers using diffusion policy imitation learning. A high-level oracle filters task-relevant information to focus each controller on a minimal observation and action space. Our graph-based neuro-symbolic framework enables capturing complex state transitions, including non-spatial and temporal relations, that data-driven learning or clustering techniques often fail to discover in limited demonstration datasets. We validate our approach in six domains that involve four robotic arms, Stacking, Kitchen, Assembly, and Towers of Hanoi environments, and a distinct Automated Forklift domain with two environments. The results demonstrate high data efficiency with as few as five skill demonstrations, strong zero- and few-shot generalizations, and interpretable decision making.

Problem

• Solves long-horizon planning with few demonstrations
• Generalizes across task variations and distribution shifts
• Learns neuro-symbolic control policies from limited data

Innovation

• Neuro-symbolic framework combining control policies and abstractions
• Graph-based task abstraction with ASP solver rule discovery
• Diffusion policy imitation learning with high-level information filtering