Hierarchical process supervision improves scene safety understanding
Combining Hierarchical Cognitive Process with Process Supervision for Interpretable Scene Safety Understanding
Computation and Language
Summary
Understanding whether a scene is safe is very important in many critical situations, but existing computer methods often can't explain how they reach their decisions. The authors created a way to teach computers to think step-by-step about safety, like humans do. They built a special dataset with detailed reasoning steps and designed a new system that breaks down safety decisions into smaller parts handled by expert modules. This new approach makes the safety decisions more understandable and works better than older methods.
What this means in practice
- •For autonomous vehicle developers: Improve safety decision-making systems with interpretable and modular reasoning on scene safety.
- •For security monitoring teams: Use detailed cognitive-style safety analysis to better assess risks in surveillance footage.
Authors
Zhiyun Jiang, Hanyong Wang, Binbin Liang, Yu Xie, Zhengjie Wang, Menglong Yang, Wei Li
Abstract
Scene safety understanding plays a life-or-death role in situational awareness in various critical domains. Traditional methods that rely on learning direct mappings between scenes and safety levels often lack interpretability, limiting their reliability in critical applications. An effective approach to overcoming this challenge lies in interpreting human cognitive processes and equipping machine models with analogous cognitive capabilities. This work explores an effective way of integrating scene safety cognitive process modeling and process supervision. Specifically, we first construct a hierarchical cognitive safety structure, which motivates the development of a novel, high-quality scene safety understanding dataset based on multi-step reasoning with process labels. This dataset serves both as a benchmark and a resource to improve the safety reasoning capabilities of Large Language Models (LLMs), while also enabling a granular analysis of intermediate reasoning steps through information flow and saliency-based techniques. Building upon this foundation, we introduce a modular and flexible process supervision framework that reflects the hierarchical nature of human cognition. This framework leverages LLMs as the core architecture and incorporates Low-Rank Adaptation(LoRA) and Mixture-of-Experts (MoE) strategies to enable specialization and collaboration among expert modules, each tasked with specific sub-processes of the overall reasoning chain. Systematic experimental evaluations and analyses confirm that our framework exhibits superior interpretability and performance characteristics compared to traditional approaches.