LiteUpdate: A Lightweight Framework for Updating AI-Generated Image Detectors

2025-11-10

AI Summary

Rapid iteration of generative AI models causes sharp performance degradation and catastrophic forgetting in existing detection methods. Method: This paper proposes a lightweight online update framework that (i) innovatively integrates confidence scores and gradient-based features to construct a boundary sample selection mechanism, and (ii) designs a multi-path fine-tuning weight fusion strategy to jointly optimize adaptation to new models and retention of prior knowledge. The framework combines three complementary update paths—pretrained initialization, representative-sample-driven adaptation, and stochastic updating—enabling efficient incremental updates without full model retraining. Contribution/Results: Evaluated on the AIDE benchmark, our method improves Midjourney image detection accuracy from 87.63% to 93.03% (+6.16%), significantly outperforming state-of-the-art continual learning and detector update approaches. It delivers a scalable, low-overhead solution for governing AI-generated content.

Technology Category

• Machine Learning: Life-Long and Continual Learning
• Computer Vision: Diffusion Models for Vision
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

The rapid progress of generative AI has led to the emergence of new generative models, while existing detection methods struggle to keep pace, resulting in significant degradation in the detection performance. This highlights the urgent need for continuously updating AI-generated image detectors to adapt to new generators. To overcome low efficiency and catastrophic forgetting in detector updates, we propose LiteUpdate, a lightweight framework for updating AI-generated image detectors. LiteUpdate employs a representative sample selection module that leverages image confidence and gradient-based discriminative features to precisely select boundary samples. This approach improves learning and detection accuracy on new distributions with limited generated images, significantly enhancing detector update efficiency. Additionally, LiteUpdate incorporates a model merging module that fuses weights from multiple fine-tuning trajectories, including pre-trained, representative, and random updates. This balances the adaptability to new generators and mitigates the catastrophic forgetting of prior knowledge. Experiments demonstrate that LiteUpdate substantially boosts detection performance in various detectors. Specifically, on AIDE, the average detection accuracy on Midjourney improved from 87.63% to 93.03%, a 6.16% relative increase.

Problem

• Updating AI-generated image detectors to adapt to new generative models efficiently
• Overcoming catastrophic forgetting during detector updates while maintaining prior knowledge
• Improving detection accuracy for new image generators with limited training samples

Innovation

• Selects boundary samples using confidence and gradient features
• Merges weights from multiple fine-tuning trajectories
• Balances new generator adaptability with prior knowledge retention