ACDC: Adaptive Curriculum Planning with Dynamic Contrastive Control for Goal-Conditioned Reinforcement Learning in Robotic Manipulation

2026-03-02

AI Summary

This work addresses the limitations of existing goal-conditioned reinforcement learning methods, which often suffer from simplistic experience replay strategies that fail to balance task diversity and learning efficiency, thereby constraining robotic manipulation performance. To overcome this, we propose the ACDC framework, which uniquely integrates multidimensional adaptive curriculum scheduling with dynamic contrastive control. ACDC adaptively guides the learning trajectory through curriculum planning and employs norm-constrained contrastive learning to enable curriculum-aware experience selection. This synergy jointly optimizes the exploration–exploitation trade-off and representation learning. Evaluated on multiple complex robotic manipulation tasks, ACDC substantially outperforms current approaches, achieving notable gains in both sample efficiency and final success rate, effectively emulating the progressive nature of human learning.

Technology Category

• Intelligent Robots: Manipulation
• Search and Optimization: Learning to Search
• Machine Learning: Adversarial Learning & Robustness

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Goal-conditioned reinforcement learning has shown considerable potential in robotic manipulation; however, existing approaches remain limited by their reliance on prioritizing collected experience, resulting in suboptimal performance across diverse tasks. Inspired by human learning behaviors, we propose a more comprehensive learning paradigm, ACDC, which integrates multidimensional Adaptive Curriculum (AC) Planning with Dynamic Contrastive (DC) Control to guide the agent along a well-designed learning trajectory. More specifically, at the planning level, the AC component schedules the learning curriculum by dynamically balancing diversity-driven exploration and quality-driven exploitation based on the agent's success rate and training progress. At the control level, the DC component implements the curriculum plan through norm-constrained contrastive learning, enabling magnitude-guided experience selection aligned with the current curriculum focus. Extensive experiments on challenging robotic manipulation tasks demonstrate that ACDC consistently outperforms the state-of-the-art baselines in both sample efficiency and final task success rate.

Problem

• goal-conditioned reinforcement learning
• robotic manipulation
• curriculum learning
• experience prioritization
• sample efficiency

Innovation

• Adaptive Curriculum Planning
• Dynamic Contrastive Control
• Goal-Conditioned Reinforcement Learning
• Norm-Constrained Contrastive Learning
• Robotic Manipulation