Decentralized Attribution of Generative Models

TL;DR

This paper proposes a decentralized approach for attributing generative models using user-specific binary classifiers, addressing scalability and guarantee issues of centralized methods, and validates it on multiple datasets.

Contribution

It introduces a decentralized attribution framework with key-based classifiers that guarantees attribution bounds, improving scalability and robustness over centralized methods.

Findings

Decentralized attribution guarantees are established under certain key conditions.

Method is validated on MNIST, CelebA, and FFHQ datasets.

Trade-offs between generation quality and robustness are analyzed.

Abstract

Growing applications of generative models have led to new threats such as malicious personation and digital copyright infringement. One solution to these threats is model attribution, i.e., the identification of user-end models where the contents under question are generated from. Existing studies showed empirical feasibility of attribution through a centralized classifier trained on all user-end models. However, this approach is not scalable in reality as the number of models ever grows. Neither does it provide an attributability guarantee. To this end, this paper studies decentralized attribution, which relies on binary classifiers associated with each user-end model. Each binary classifier is parameterized by a user-specific key and distinguishes its associated model distribution from the authentic data distribution. We develop sufficient conditions of the keys that guarantee an attributability lower bound. Our method is validated on MNIST, CelebA, and FFHQ datasets. We also examine the trade-off between generation quality and robustness of attribution against adversarial post-processes.