Predictability-Based Curiosity-Guided Action Symbol Discovery

2025-05-23

AI Summary

This work addresses developmental robotics by autonomously discovering dual symbolic representations—both action and perceptual—without human-provided priors, to enable abstract reasoning and goal-directed planning. Method: We propose a prediction-driven joint symbolic discovery framework: an encoder-decoder network models action effects, and effect-distribution entropy defines a curiosity reward that guides active exploration of high-information actions in continuous action spaces. Crucially, we integrate predictability modeling with curiosity for the first time to jointly generate action and perceptual symbols, and construct a symbolic tree-search planner. Results: Experiments on single- and dual-object manipulation tasks demonstrate successful acquisition of semantically coherent and diverse action symbols. Our approach significantly improves planning success rates, outperforming both random exploration and pure reinforcement learning baselines.

Technology Category

• Humans and AI: Human-Aware Planning and Behavior Prediction
• Computer Vision: Visual Reasoning & Symbolic Representations
• Cognitive Modeling & Cognitive Systems: Symbolic Representations

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• 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

Abstract

Discovering symbolic representations for skills is essential for abstract reasoning and efficient planning in robotics. Previous neuro-symbolic robotic studies mostly focused on discovering perceptual symbolic categories given a pre-defined action repertoire and generating plans with given action symbols. A truly developmental robotic system, on the other hand, should be able to discover all the abstractions required for the planning system with minimal human intervention. In this study, we propose a novel system that is designed to discover symbolic action primitives along with perceptual symbols autonomously. Our system is based on an encoder-decoder structure that takes object and action information as input and predicts the generated effect. To efficiently explore the vast continuous action parameter space, we introduce a Curiosity-Based exploration module that selects the most informative actions -- the ones that maximize the entropy in the predicted effect distribution. The discovered symbolic action primitives are then used to make plans using a symbolic tree search strategy in single- and double-object manipulation tasks. We compare our model with two baselines that use different exploration strategies in different experiments. The results show that our approach can learn a diverse set of symbolic action primitives, which are effective for generating plans in order to achieve given manipulation goals.

Problem

• Discover symbolic action primitives autonomously for robotics
• Explore action space using curiosity-driven entropy maximization
• Enable abstract planning in object manipulation tasks

Innovation

• Autonomous discovery of symbolic action primitives
• Curiosity-Based exploration maximizes effect entropy
• Encoder-decoder predicts effects from actions