EvoKnow: Continual Knowledge Evolution for AI-Generated Image Detection

Key points

• AI-generated image detectors are commonly trained on fixed generator domains and become difficult to maintain as new generative models emerge.1
• Continual adaptation is challenging because replaying historical generated images is costly, whereas updating shared parameters with limited current-domain data can overwrite prior forensic knowledge.1
• EvoKnow preserves a shared forensic basis learned from base domains, incrementally adds isolated residual experts for complementary generator-relevant evidence, and retrieves expertise through an Analytical Incremental Router (AIR) updated in closed form from current-stage generated images and accumulated sufficient statistics.1
• Experiments demonstrate effective cross-generator generalization, few-shot expansion, and long-horizon continual adaptation.1

Sources (1)

• [1]EvoKnow: Continual Knowledge Evolution for AI-Generated Image Detection
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 8, 07:12 AM

  We propose EvoKnow, a replay-free framework that formulates continual AI-generated image detection as forensic knowledge evolution.

  AI-generated image detectors are commonly trained on fixed generator domains and become difficult to maintain as new generative models emerge.

Extractive summary: sentences quoted from the sources.