Fake-HR1: Rethinking Reasoning of Vision Language Model for Synthetic Image Detection

2026-02-10

AI Summary

This work addresses the inefficiency of existing vision-language models in synthetic image detection, which rely on fixed chain-of-thought reasoning and perform redundant computations even on obviously forged samples. To overcome this limitation, the authors propose Fake-HR1, the first approach to introduce adaptive reasoning into generated image detection. Fake-HR1 employs a two-stage training framework: it begins with hybrid fine-tuning (HFT) for cold-start initialization, followed by online reinforcement learning based on a hierarchical grouped reward policy optimization (HGRPO) to dynamically decide whether to invoke complex reasoning. By adaptively adjusting the inference path according to input image characteristics, Fake-HR1 achieves significant improvements over current large language models across diverse query types, setting new benchmarks in both detection accuracy and response efficiency.

Technology Category

• Computer Vision: Large Vision Models
• Natural Language Processing: Fact-Checking / Misinformation Detection (NLP Focus)
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Recent studies have demonstrated that incorporating Chain-of-Thought (CoT) reasoning into the detection process can enhance a model's ability to detect synthetic images. However, excessively lengthy reasoning incurs substantial resource overhead, including token consumption and latency, which is particularly redundant when handling obviously generated forgeries. To address this issue, we propose Fake-HR1, a large-scale hybrid-reasoning model that, to the best of our knowledge, is the first to adaptively determine whether reasoning is necessary based on the characteristics of the generative detection task. To achieve this, we design a two-stage training framework: we first perform Hybrid Fine-Tuning (HFT) for cold-start initialization, followed by online reinforcement learning with Hybrid-Reasoning Grouped Policy Optimization (HGRPO) to implicitly learn when to select an appropriate reasoning mode. Experimental results show that Fake-HR1 adaptively performs reasoning across different types of queries, surpassing existing LLMs in both reasoning ability and generative detection performance, while significantly improving response efficiency.

Problem

• synthetic image detection
• Chain-of-Thought reasoning
• resource overhead
• visual language model
• reasoning efficiency

Innovation

• adaptive reasoning
• synthetic image detection
• hybrid reasoning
• Chain-of-Thought
• reinforcement learning