A Rich Knowledge Space for Scalable Deepfake Detection

ICLR 2026 Conference Submission11645 Authors

Keywords: Deepfake Detection, Media Forensics, Multi-modal Learning

TL;DR: Harnessing large-scale pre-training for robust deepfake detection.

Abstract: The proliferation of realistic deepfakes has driven the development of numerous benchmark datasets to support detection research. Despite their increasing volume and diversity, no prior effort has systematically consolidated these resources into a unified framework for large-scale model training, nor has there been a massively pre-trained model tailored to deepfake detection. In this work, we introduce MMI-DD (Multi-modal Multi-type Integrated Deepfake Dataset), a large-scale resource containing 3.6 million facial images, the largest collection to date. It unifies diverse benchmarks with uniform preprocessing, and further provides fine-grained annotations across four deepfake types, as well as VLM-generated descriptions capturing both facial and environmental attributes for each image. By leveraging this comprehensive multi-modal dataset, we construct a foundational deepfake knowledge space that empowers our model to discern a broad spectrum of synthetic media. Our method, SD^2 (Scalable Deepfake Detection), refines CLIP for deepfake detection, optimizing image-text classification with rich, type-specific labels. We enhance this with intermediate visual features capturing low-level cues and text label separation loss for stability. We further leverage VLM-generated descriptions and contrastive learning to expand the scope of forgery knowledge, reducing overfitting and enhancing generalization. Extensive experiments on challenging deepfake datasets and AIGC benchmark demonstrate the effectiveness, scalability, and real-world applicability of our approach. Our dataset and code will be available at https://anonymous.4open.science/r/SDD/.

Primary Area: alignment, fairness, safety, privacy, and societal considerations

Submission Number: 11645