Invariant Discovery for Networked Systems

2026-07-24

AI Summary

This work addresses the challenge of manually specifying invariants among signals in networked systems, which demands deep expertise in formal logic and domain knowledge, while existing automated approaches are constrained by predefined syntactic templates and struggle with noise in real-world telemetry data. To overcome these limitations, the paper proposes a two-stage framework: first leveraging the semantic reasoning capabilities of large language models to automatically discover candidate invariant grammars, then applying a statistical rule-search procedure coupled with formal verification to ensure reliability and auditability. This approach uniquely integrates large language models with statistical validation, eliminating dependence on handcrafted templates and enabling efficient extraction of formally guaranteed network invariants directly from open telemetry streams. Evaluated on both public and production datasets, the method successfully reproduces expert-defined invariants with high coverage and low false-positive rates.

Abstract

Invariants, the relations expected to hold among measured signals of a network, underpin applications from verification to traffic generation, telemetry imputation, and input validation, yet writing them by hand demands rare expertise in both formal logic and networking. Automatic miners can help but fall short on two fronts: they still require the hardest input (the grammar of admissible invariants) and they learn only exact, ``hard'' rules, struggling with real-world approximation caused by inherent noise in data. LLMs are tools that can provide semantic reasoning over data, but are non-deterministic and opaque in their learning. Our key idea is to partition the invariant search problem into an AI-driven grammar ``discovery'' problem, followed by a statistics-driven ``search'' problem within the learned grammar. Taken together, this allows non-deterministic, hallucination-prone AI to help produce auditable invariants with formal guarantees. We design and implement such a system, Autogram, and evaluate it on both public and production telemetry data, recovering expert-derived invariants with high coverage and low false positives. We close with discussion on open problems on the path toward fully open-ended discovery.

Problem

• invariant discovery
• networked systems
• grammar learning
• noisy data
• formal guarantees

Innovation

• invariant discovery
• grammar learning
• networked systems
• large language models
• statistical verification