The FABRIC Strategy for Verifying Neural Feedback Systems

2026-03-09

AI Summary

This work addresses the scalability bottleneck in existing neural feedback systems for backward reachability analysis, which limits their ability to perform safety verification. The paper proposes a novel algorithm that integrates forward and backward reachability analysis, introducing—for the first time—a systematic and scalable approach to compute both over- and under-approximations of backward reachable sets. This method is effectively combined with forward analysis to verify reach-avoid properties of nonlinear neural feedback systems. Built upon this algorithm, the FaBRIC verification framework demonstrates significant improvements over state-of-the-art methods across multiple benchmarks, achieving superior accuracy and computational efficiency.

Technology Category

• Natural Language Processing: Safety and Robustness
• Multiagent Systems: Adversarial Agents
• Machine Learning: Calibration & Uncertainty Quantification

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Systems and Infrastructure for Web, Mobile and WoT: Experiences and lessons learnt from Web-based algorithms and system deployments
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Forward reachability analysis is a dominant approach for verifying reach-avoid specifications in neural feedback systems, i.e., dynamical systems controlled by neural networks, and a number of directions have been proposed and studied. In contrast, far less attention has been given to backward reachability analysis for these systems, in part because of the limited scalability of known techniques. In this work, we begin to address this gap by introducing new algorithms for computing both over- and underapproximations of backward reachable sets for nonlinear neural feedback systems. We also describe and implement an integration of these backward reachability techniques with existing ones for forward analysis. We call the resulting algorithm Forward and Backward Reachability Integration for Certification (FaBRIC). We evaluate our algorithms on a representative set of benchmarks and show that they significantly outperform the prior state of the art.

Problem

• neural feedback systems
• backward reachability analysis
• reach-avoid specifications
• scalability
• verification

Innovation

• backward reachability
• neural feedback systems
• reachability analysis
• FaBRIC
• formal verification