Model Fingerprinting with Benign Inputs - Inria - Institut national de recherche en sciences et technologies du numérique Accéder directement au contenu
Communication Dans Un Congrès Année : 2023

Model Fingerprinting with Benign Inputs

Résumé

Recent advances in the fingerprinting of deep neural networks are able to detect specific instances of models, placed in a black-box interaction scheme. Inputs used by the fingerprinting protocols are specifically crafted for each precise model to be checked for. While efficient in such a scenario, this nevertheless results in a lack of guarantee after a mere modification of a model (e.g. finetuning, quantization of the parameters). In this paper we propose fingerprinting scheme (coined FBI) that are resilient to significant modifications of the models. These modifications are viewed and modeled as variants. We demonstrate that benign inputs, that are unmodified images, are sufficient material for efficient fingerprinting. We leverage an information-theoretic approach to achieve a success rate of 95.2%. It is experimentally validated over an unprecedented set of more than 1,000 neural networks, while demonstrating performance improvements over a state-of-the-art fingerprinting method.
Fichier principal
Vignette du fichier
main.pdf (274.64 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04112859 , version 1 (01-06-2023)

Licence

Paternité

Identifiants

Citer

Thibault Maho, Teddy Furon, Erwan Le Merrer. Model Fingerprinting with Benign Inputs. ICASSP 2023 - IEEE International Conference on Acoustics, Speech and Signal Processing, Jun 2023, Ialyssos, Greece. pp.1-4, ⟨10.1109/ICASSP49357.2023.10094751⟩. ⟨hal-04112859⟩
32 Consultations
42 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More