DNA: Uncovering Universal Latent Forgery Knowledge

TL;DR

The paper introduces discriminative neural anchors (DNA), a coarse-to-fine probing framework that awakens latent forgery-detection knowledge stored in pre-trained backbone neurons. By localizing critical intermediate layers and extracting a sparse set of forgery-discriminative units (FDUs) via a triadic fusion score and curvature-based pruning, DNA achieves strong few-shot generalization and robustness across architectures and unseen generative models. The authors validate their approach on a new high-fidelity benchmark, HIFI-Gen, and demonstrate superior performance and efficiency compared to state-of-the-art methods, including notable improvements on challenging models like Midjourney. This work suggests that intrinsic representations learned during large-scale pre-training can serve as a universal, data-efficient defense against hyper-realistic AI-generated content, without extensive fine-tuning.

Abstract

As generative AI achieves hyper-realism, superficial artifact detection has become obsolete. While prevailing methods rely on resource-intensive fine-tuning of black-box backbones, we propose that forgery detection capability is already encoded within pre-trained models rather than requiring end-to-end retraining. To elicit this intrinsic capability, we propose the discriminative neural anchors (DNA) framework, which employs a coarse-to-fine excavation mechanism. First, by analyzing feature decoupling and attention distribution shifts, we pinpoint critical intermediate layers where the focus of the model logically transitions from global semantics to local anomalies. Subsequently, we introduce a triadic fusion scoring metric paired with a curvature-truncation strategy to strip away semantic redundancy, precisely isolating the forgery-discriminative units (FDUs) inherently imprinted with sensitivity to forgery traces. Moreover, we introduce HIFI-Gen, a high-fidelity synthetic benchmark built upon the very latest models, to address the lag in existing datasets. Experiments demonstrate that by solely relying on these anchors, DNA achieves superior detection performance even under few-shot conditions. Furthermore, it exhibits remarkable robustness across diverse architectures and against unseen generative models, validating that waking up latent neurons is more effective than extensive fine-tuning.

Figures (16)

• Figure 1: Comparison of detection paradigms in the era of hyper-realistic generation. Unlike conventional methods that fail as surface artifacts disappear, our DNA framework exploits neuronal hierarchies to robustly detect hyper-realistic scenarios.
• Figure 2: Visualization of HIFI-Gen structure. HIFI-Gen comprises images generated by five distinct generative models, each yielding 3,000 images across 1,000 categories.
• Figure 3: Overall of the DNA framework. The pipeline operates in a coarse-to-fine manner. Stage 1: layer localization. We pinpoint critical intermediate layers by analyzing feature-space decoupling and attention-distribution shifts, validated by linear probing. Stage 2: FDUs detection. We identify sparse forgery-discriminative units (FDUs) from the frozen backbone using a triadic fusion score ($S_{i,k}$).
• Figure 4: Visualization of distance metrics across layers. The cosine distance ($D_{cos}$) between the centroids of real and fake classes at each layer. The euclidean distance ($D_{L2}$) of the global attention distributions between real and fake images.
• Figure 5: Visualization of FDUs attention. The activation maps of FDUs on both original and augmented images.