Generative Artificial Intelligence Reproducibility and Consensus

TL;DR

This paper investigates the reproducibility and verification of generative AI outputs using large-scale comparisons, consensus methods, and empirical analysis to enhance trust, transparency, and scientific rigor in AI research.

Contribution

It introduces a comprehensive methodology for verifying generative AI results through large-scale comparisons, consensus techniques, and analysis of stochasticity sources, advancing reproducibility standards.

Findings

Over 99.89% probability of detecting perceptual collisions in images

Achieved 100% consensus in large language models using greedy and beam search methods

Provided empirical bounds for verification error and stochasticity sources

Abstract

We performed a billion locality sensitive hash comparisons between artificially generated data samples to answer the critical question - can we reproduce the results of generative AI models? Reproducibility is one of the pillars of scientific research for verifiability, benchmarking, trust, and transparency. Futhermore, we take this research to the next level by verifying the "correctness" of generative AI output in a non-deterministic, trustless, decentralized network. We generate millions of data samples from a variety of open source diffusion and large language models and describe the procedures and trade-offs between generating more verses less deterministic output. Additionally, we analyze the outputs to provide empirical evidence of different parameterizations of tolerance and error bounds for verification. For our results, we show that with a majority vote between three independent verifiers, we can detect image generated perceptual collisions in generated AI with over 99.89% probability and less than 0.0267% chance of intra-class collision. For large language models (LLMs), we are able to gain 100% consensus using greedy methods or n-way beam searches to generate consensus demonstrated on different LLMs. In the context of generative AI training, we pinpoint and minimize the major sources of stochasticity and present gossip and synchronization training techniques for verifiability. Thus, this work provides a practical, solid foundation for AI verification, reproducibility, and consensus for generative AI applications.