LTL Verification of Memoryful Neural Agents

2025-03-04

AI Summary

This paper addresses the formal verification of full Linear Temporal Logic (LTL) specifications for memory-augmented neural multi-agent systems (MN-MAS) operating in non-deterministic, partially observable environments—introducing the first framework supporting *unbounded* LTL properties. Methodologically, it integrates lasso-based search with invariant synthesis to overcome the fundamental limitation of prior approaches, which are restricted to bounded verification. To enhance scalability, the framework synergistically combines bounded model checking, interval propagation, mixed-integer linear programming, and adaptive partitioning for constraint solving. Experimental evaluation on Gymnasium and PettingZoo benchmarks demonstrates that our approach achieves the first successful unbounded LTL verification for MN-MAS. Moreover, for bounded verification, it outperforms state-of-the-art methods by an order of magnitude in efficiency.

Technology Category

• Multiagent Systems: Multiagent Learning
• Machine Learning: Large Multimodal Models (LMMs)
• Constraint Satisfaction and Optimization: Satisfiability Modulo Theories

Application Category

• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

We present a framework for verifying Memoryful Neural Multi-Agent Systems (MN-MAS) against full Linear Temporal Logic (LTL) specifications. In MN-MAS, agents interact with a non-deterministic, partially observable environment. Examples of MN-MAS include multi-agent systems based on feed-forward and recurrent neural networks or state-space models. Different from previous approaches, we support the verification of both bounded and unbounded LTL specifications. We leverage well-established bounded model checking techniques, including lasso search and invariant synthesis, to reduce the verification problem to that of constraint solving. To solve these constraints, we develop efficient methods based on bound propagation, mixed-integer linear programming, and adaptive splitting. We evaluate the effectiveness of our algorithms in single and multi-agent environments from the Gymnasium and PettingZoo libraries, verifying unbounded specifications for the first time and improving the verification time for bounded specifications by an order of magnitude compared to the SoA.

Problem

• Verification of Memoryful Neural Multi-Agent Systems
• Support for bounded and unbounded LTL specifications
• Efficient constraint solving using advanced techniques

Innovation

• Verifies Memoryful Neural Multi-Agent Systems
• Uses bounded model checking techniques
• Develops efficient constraint solving methods