Curiosity-Driven Testing for Sequential Decision-Making Process

2025-09-02

AI Summary

To address the susceptibility of safety-critical sequential decision-making systems to learning unsafe behaviors, this paper proposes a curiosity-driven black-box fuzzing framework for efficiently discovering diverse crash-triggering scenarios. Methodologically, it innovatively integrates an intrinsic curiosity mechanism with a multi-objective seed selection strategy to dynamically balance exploration of novel states and fault triggering. Technically, it incorporates deep learning–based mutation generation, prediction-uncertainty–guided novelty measurement, and Pareto-optimal seed scheduling. Experimental evaluation across multiple mainstream sequential decision models demonstrates that the approach significantly outperforms existing state-of-the-art methods: it achieves a 23.6% improvement in fault detection rate and a 41.2% increase in crash-scenario diversity. Moreover, the generated diverse failure cases facilitate subsequent model repair and robustness enhancement.

Technology Category

• Search and Optimization: Sampling/Simulation-based Search
• Reasoning under Uncertainty: Sequential Decision Making
• Constraint Satisfaction and Optimization: Constraint Learning and Acquisition

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Security and Privacy: Large-scale security measurements

Abstract

Sequential decision-making processes (SDPs) are fundamental for complex real-world challenges, such as autonomous driving, robotic control, and traffic management. While recent advances in Deep Learning (DL) have led to mature solutions for solving these complex problems, SDMs remain vulnerable to learning unsafe behaviors, posing significant risks in safety-critical applications. However, developing a testing framework for SDMs that can identify a diverse set of crash-triggering scenarios remains an open challenge. To address this, we propose CureFuzz, a novel curiosity-driven black-box fuzz testing approach for SDMs. CureFuzz proposes a curiosity mechanism that allows a fuzzer to effectively explore novel and diverse scenarios, leading to improved detection of crashtriggering scenarios. Additionally, we introduce a multi-objective seed selection technique to balance the exploration of novel scenarios and the generation of crash-triggering scenarios, thereby optimizing the fuzzing process. We evaluate CureFuzz on various SDMs and experimental results demonstrate that CureFuzz outperforms the state-of-the-art method by a substantial margin in the total number of faults and distinct types of crash-triggering scenarios. We also demonstrate that the crash-triggering scenarios found by CureFuzz can repair SDMs, highlighting CureFuzz as a valuable tool for testing SDMs and optimizing their performance.

Problem

• Testing sequential decision-making processes for unsafe behaviors
• Identifying diverse crash-triggering scenarios in safety-critical applications
• Developing effective fuzz testing framework for deep learning systems

Innovation

• Curiosity-driven black-box fuzz testing approach
• Multi-objective seed selection technique balancing exploration
• Novel curiosity mechanism exploring diverse crash scenarios