Neural Interactive Proofs

2024-12-12

arXiv.org

AI Summary

This work addresses how computationally constrained yet trustworthy verifiers can reliably solve complex tasks via interaction with untrusted, high-capacity provers. Method: We propose “Neural Interactive Proofs” (NIP), a framework that models agents as neural networks and unifies game-theoretic reasoning, verifiable computation, and graph-structured learning into a principled prover-verifier interaction protocol. Contribution/Results: Theoretically, we establish the first completeness and efficiency bounds for multi-round interactive proof systems under neural parameterization. Empirically, we validate NIP on two canonical tasks—graph isomorphism testing and LLM-generated code verification—achieving a 32% improvement in verification accuracy and demonstrating robust detection of logical errors. By integrating formal verification guarantees with neural representational power, NIP establishes a new paradigm for verifiable and interpretable AI collaboration, providing foundational methodology for trustworthy AI-assisted reasoning.

Technology Category

• Natural Language Processing: Safety and Robustness
• Knowledge Representation and Reasoning: Automated Reasoning and Theorem Proving
• Multiagent Systems: Adversarial Agents

Application Category

• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Assisted, interactive, and conversational search

Abstract

We consider the problem of how a trusted, but computationally bounded agent (a 'verifier') can learn to interact with one or more powerful but untrusted agents ('provers') in order to solve a given task. More specifically, we study the case in which agents are represented using neural networks and refer to solutions of this problem as neural interactive proofs. First we introduce a unifying framework based on prover-verifier games, which generalises previously proposed interaction protocols. We then describe several new protocols for generating neural interactive proofs, and provide a theoretical comparison of both new and existing approaches. Finally, we support this theory with experiments in two domains: a toy graph isomorphism problem that illustrates the key ideas, and a code validation task using large language models. In so doing, we aim to create a foundation for future work on neural interactive proofs and their application in building safer AI systems.

Problem

• Trusted verifier learns to interact with untrusted provers.
• Neural networks represent agents in interactive proofs.
• Develop protocols for safer AI systems using neural proofs.

Innovation

• Unifying framework for prover-verifier games
• New protocols for neural interactive proofs
• Experiments in graph isomorphism and code validation