Continuous fake media detection: adapting deepfake detectors to new generative techniques

View on arXiv ← Back to list

Authors: Francesco Tassone, Luca Maiano, Irene Amerini

Published: 2024-06-12 13:04:06+00:00

AI Summary

This paper analyzes the application of continual learning (specifically Knowledge Distillation and Elastic Weight Consolidation) to improve the robustness of deepfake detectors against evolving generative techniques. The authors demonstrate that continual learning helps maintain performance across sequences of deepfakes, but performance is significantly impacted by task similarity and order.

Abstract

Generative techniques continue to evolve at an impressively high rate, driven by the hype about these technologies. This rapid advancement severely limits the application of deepfake detectors, which, despite numerous efforts by the scientific community, struggle to achieve sufficiently robust performance against the ever-changing content. To address these limitations, in this paper, we propose an analysis of two continuous learning techniques on a Short and a Long sequence of fake media. Both sequences include a complex and heterogeneous range of deepfakes generated from GANs, computer graphics techniques, and unknown sources. Our study shows that continual learning could be important in mitigating the need for generalizability. In fact, we show that, although with some limitations, continual learning methods help to maintain good performance across the entire training sequence. For these techniques to work in a sufficiently robust way, however, it is necessary that the tasks in the sequence share similarities. In fact, according to our experiments, the order and similarity of the tasks can affect the performance of the models over time. To address this problem, we show that it is possible to group tasks based on their similarity. This small measure allows for a significant improvement even in longer sequences. This result suggests that continual techniques can be combined with the most promising detection methods, allowing them to catch up with the latest generative techniques. In addition to this, we propose an overview of how this learning approach can be integrated into a deepfake detection pipeline for continuous integration and continuous deployment (CI/CD). This allows you to keep track of different funds, such as social networks, new generative tools, or third-party datasets, and through the integration of continuous learning, allows constant maintenance of the detectors.

Key findings

Continual learning methods (particularly Knowledge Distillation) outperformed transfer learning in maintaining deepfake detection accuracy across sequences. However, performance was sensitive to task similarity and order within the sequences. Grouping similar tasks improved overall performance, suggesting that task similarity is crucial for effective continual learning in this context.

Approach

The authors address the problem of deepfake detector adaptation to new generative techniques by employing continual learning methods (Knowledge Distillation and Elastic Weight Consolidation). They evaluate these methods on sequences of deepfake data, examining the impact of task order and similarity on model performance.

Datasets

CDDB dataset, including easy, hard, and long sequences of deepfakes from various GANs, non-GAN models, and unknown sources.

Model(s)

ResNet-50, ResNet-18, MobileNet-V2, and Xception.

Author countries

Italy

← Previous