为多模块AI系统提供细粒度故障归因，提升复杂模型的鲁棒性测试能力
SETA: Statistical Fault Attribution for Compound AI Systems
- 分模块测试框架，可定位错误来源
- 支持跨模块错误传播分析，突破端到端评估局限
- 适用于多模态、跨领域的多网络系统，真实场景验证有效
现代AI系统越来越多地由多个相互连接的神经网络组成，以处理复杂的推理任务。对这类系统进行鲁棒性和安全性测试面临巨大挑战。现有的先进鲁棒性测试技术（无论是黑盒还是白盒）主要针对单个网络模型设计，难以扩展至多网络流水线。本文提出一种模块化鲁棒性测试框架，对测试数据施加一组扰动，支持（1）组件级系统分析以隔离错误，（2）对神经网络模块间错误传播进行推理。该框架具有架构和模态无关性，可跨领域应用。我们将框架应用于一个由多个深度网络组成的实际自动驾驶轨道检测系统，成功展示了该方法如何实现超越传统端到端指标的细粒度鲁棒性分析。
原文摘要 · Abstract (English)
Modern AI systems increasingly comprise multiple interconnected neural networks to tackle complex inference tasks. Testing such systems for robustness and safety entails significant challenges. Current state-of-the-art robustness testing techniques, whether black-box or white-box, have been proposed and implemented for single-network models and do not scale well to multi-network pipelines. We propose a modular robustness testing framework that applies a given set of perturbations to test data. Our testing framework supports (1) a component-wise system analysis to isolate errors and (2) reasoning about error propagation across the neural network modules. The testing framework is architecture and modality agnostic and can be applied across domains. We apply the framework to a real-world autonomous rail inspection system composed of multiple deep networks and successfully demonstrate how our approach enables fine-grained robustness analysis beyond conventional end-to-end metrics.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。