On the Reliability of Networks of AI Agents: Density Evolution, Stopping Sets, and Architecture Optimization

2026-06-16

AI Summary

This work addresses the lack of theoretical characterization of success mechanisms and reliability under multiple failure modes in multi-agent AI systems performing collaborative tasks. The authors model collaboration as message passing on a sparse, role-structured factor graph, introducing a set-valued propagation scheme that incorporates three types of erasure-like failures and integrates nonlinear, value-asymmetric logical verifiers. They extend density evolution theory—previously limited to linear settings—to agent networks featuring nonlinear verification and multiple failure modes, revealing an inherent asymmetry between positive and negative verification under logical operators such as AND. A new threshold and finite-length analysis framework is established for both deterministic and random graph sequences. The theory accurately predicts the asymptotic fraction of unresolved sub-propositions, recovers classical LDPC-BEC results in the XOR case, and uncovers asymmetric behavior of verification certificates in the AND case.

Abstract

Modern AI systems increasingly solve a task not with a single model call but with several imperfect agents working together: some propose pieces of a solution, others verify them, and the results are combined. These systems often outperform any single model, yet it is rarely clear why they succeed or when they will fail. We model such a system as message passing on a sparse graph, the structure that underlies low-density parity-check (LDPC) codes, and extend the density-evolution machinery of coding theory to this richer setting. In our model a task is a set of coupled binary subclaims, and an agent architecture is a sparse, role-typed factor graph whose check nodes are noisy Boolean verifier nodes, each computing a local Boolean function of the subclaims it touches. Three distinct failure modes, all modeled as erasures (an agent abstaining, a verifier returning no usable output, and a message lost between two agents), propagate as the agents exchange set-valued messages. The check agents combine these messages by a single logical-forcing rule that specializes to XOR, AND, OR, implication, and Horn constraints. This is more than a relabeling of LDPC theory: the verifier functions are nonlinear and value-asymmetric, and the three failure modes do not reduce to a single effective channel, so they require new threshold, finite-length, and converse results rather than a direct reuse of parity-check density evolution. We prove a density-evolution theorem that predicts the asymptotic fraction of unresolved subclaims on random role-typed architectures, with an extension to deterministic, locally tree-like graph sequences. The XOR case recovers the classical LDPC recursion on the binary erasure channel (BEC); the AND case exposes an asymmetry between positive and negative verifier certificates.

Problem

• AI agent networks
• reliability
• failure modes
• message passing
• factor graphs

Innovation

• density evolution
• AI agent networks
• factor graphs
• Boolean verifiers
• stopping sets