Straight Through Gumbel Softmax Estimator based Bimodal Neural Architecture Search for Audio-Visual Deepfake Detection

TL;DR

This paper introduces a novel neural architecture search framework using Straight-through Gumbel-Softmax for multimodal deepfake detection, significantly improving fusion model performance and robustness.

Contribution

It proposes a comprehensive two-level search approach for optimizing multimodal fusion architectures specifically for deepfake detection tasks.

Findings

Achieved 94.4% AUC on FakeAVCeleb and SWAN-DF datasets.

Efficiently identified crucial features from backbone networks.

Developed a fusion architecture with minimal model parameters.

Abstract

Deepfakes are a major security risk for biometric authentication. This technology creates realistic fake videos that can impersonate real people, fooling systems that rely on facial features and voice patterns for identification. Existing multimodal deepfake detectors rely on conventional fusion methods, such as majority rule and ensemble voting, which often struggle to adapt to changing data characteristics and complex patterns. In this paper, we introduce the Straight-through Gumbel-Softmax (STGS) framework, offering a comprehensive approach to search multimodal fusion model architectures. Using a two-level search approach, the framework optimizes the network architecture, parameters, and performance. Initially, crucial features were efficiently identified from backbone networks, whereas within the cell structure, a weighted fusion operation integrated information from various sources. An architecture that maximizes the classification performance is derived by varying parameters such as temperature and sampling time. The experimental results on the FakeAVCeleb and SWAN-DF datasets demonstrated an impressive AUC value 94.4\% achieved with minimal model parameters.