Zhonghao Zhang

Wayne Zhang, Changjiang Jiang, Zhonghao Zhang et al.

The rapid advancement of Artificial Intelligence Generated Content (AIGC) in visual domains has resulted in highly realistic synthetic images and videos, driven by sophisticated generative frameworks such as diffusion-based architectures. While these breakthroughs open substantial opportunities, they simultaneously raise critical concerns about content authenticity and integrity. Many current AIGC detection methods operate as black-box binary classifiers, which offer limited interpretability, and no approach supports detecting both images and videos in a unified framework. This dual limitation compromises model transparency, reduces trustworthiness, and hinders practical deployment. To address these challenges, we introduce IVY-FAKE , a novel, unified, and large-scale dataset specifically designed for explainable multimodal AIGC detection. Unlike prior benchmarks, which suffer from fragmented modality coverage and sparse annotations, IVY-FAKE contains over 150,000 richly annotated training samples (images and videos) and 18,700 evaluation examples, each accompanied by detailed natural-language reasoning beyond simple binary labels. Building on this, we propose Ivy Explainable Detector (IVY-XDETECTOR), a unified AIGC detection and explainable architecture that jointly performs explainable detection for both image and video content. Our unified vision-language model achieves state-of-the-art performance across multiple image and video detection benchmarks, highlighting the significant advancements enabled by our dataset and modeling framework. Our data is publicly available at https://huggingface.co/datasets/AI-Safeguard/Ivy-Fake.

Jia Wei, Zhonghao Zhang, Ping Chen et al.

Low-Rank Adaptation (LoRA) dominates parameter-efficient fine-tuning of large language models, yet most variants target dense architectures. Mixture-of-Experts (MoE) models scale parameters at near-constant per-token compute, and their sparse activation patterns create untapped opportunities for more efficient adaptation. We propose Hot-Experts Layer-level Low-Rank Adaptation (HELLoRA), which attaches LoRA modules only to the most frequently activated experts at each layer. This simple mechanism reduces trainable parameters and adapter-induced FLOPs while improving downstream performance, an effect we attribute to a form of structured regularization that preserves pretrained expert specialization. To stress-test HELLoRA under extreme parameter budgets, we further compose it with LoRI to form HELLoRI, which freezes the up-projection and sparsifies the down-projection. Across three MoE backbones, namely OlMoE-1B-7B, Mixtral-8x7B, and DeepSeekMoE, and three task families covering mathematical reasoning, code generation, and safety alignment, HELLoRA consistently outperforms strong PEFT baselines. Relative to vanilla LoRA on OlMoE, HELLoRA uses 15.7% of the trainable parameters, reduces adapter FLOPs by 38.7%, achieves 1.9x the training throughput, and improves accuracy by 9.2%. On DeepSeekMoE, HELLoRA outperforms LoRA while using only 23.2% of its trainable parameters. These results demonstrate that activation-aware adapter placement is an effective and practical route to scaling PEFT for MoE language models.

Ming Zhao, Wenhui Dong, Yang Zhang et al.

Spine disorders affect 619 million people globally and are a leading cause of disability, yet AI-assisted diagnosis remains limited by the lack of level-aware, multimodal datasets. Clinical decision-making for spine disorders requires sophisticated reasoning across X-ray, CT, and MRI at specific vertebral levels. However, progress has been constrained by the absence of traceable, clinically-grounded instruction data and standardized, spine-specific benchmarks. To address this, we introduce SpineMed, an ecosystem co-designed with practicing spine surgeons. It features SpineMed-450k, the first large-scale dataset explicitly designed for vertebral-level reasoning across imaging modalities with over 450,000 instruction instances, and SpineBench, a clinically-grounded evaluation framework. SpineMed-450k is curated from diverse sources, including textbooks, guidelines, open datasets, and ~1,000 de-identified hospital cases, using a clinician-in-the-loop pipeline with a two-stage LLM generation method (draft and revision) to ensure high-quality, traceable data for question-answering, multi-turn consultations, and report generation. SpineBench evaluates models on clinically salient axes, including level identification, pathology assessment, and surgical planning. Our comprehensive evaluation of several recently advanced large vision-language models (LVLMs) on SpineBench reveals systematic weaknesses in fine-grained, level-specific reasoning. In contrast, our model fine-tuned on SpineMed-450k demonstrates consistent and significant improvements across all tasks. Clinician assessments confirm the diagnostic clarity and practical utility of our model's outputs.

Zhonghao Zhang, Yipeng Liu, Xingyu Cao et al.

Deep learning has been used to image compressive sensing (CS) for enhanced reconstruction performance. However, most existing deep learning methods train different models for different subsampling ratios, which brings additional hardware burden. In this paper, we develop a general framework named scalable deep compressive sensing (SDCS) for the scalable sampling and reconstruction (SSR) of all existing end-to-end-trained models. In the proposed way, images are measured and initialized linearly. Two sampling masks are introduced to flexibly control the subsampling ratios used in sampling and reconstruction, respectively. To make the reconstruction model adapt to any subsampling ratio, a training strategy dubbed scalable training is developed. In scalable training, the model is trained with the sampling matrix and the initialization matrix at various subsampling ratios by integrating different sampling matrix masks. Experimental results show that models with SDCS can achieve SSR without changing their structure while maintaining good performance, and SDCS outperforms other SSR methods.

Zhonghao Zhang, Yipeng Liu, Jiani Liu et al.

Most compressive sensing (CS) reconstruction methods can be divided into two categories, i.e. model-based methods and classical deep network methods. By unfolding the iterative optimization algorithm for model-based methods onto networks, deep unfolding methods have the good interpretation of model-based methods and the high speed of classical deep network methods. In this paper, to solve the visual image CS problem, we propose a deep unfolding model dubbed AMP-Net. Rather than learning regularization terms, it is established by unfolding the iterative denoising process of the well-known approximate message passing algorithm. Furthermore, AMP-Net integrates deblocking modules in order to eliminate the blocking artifacts that usually appear in CS of visual images. In addition, the sampling matrix is jointly trained with other network parameters to enhance the reconstruction performance. Experimental results show that the proposed AMP-Net has better reconstruction accuracy than other state-of-the-art methods with high reconstruction speed and a small number of network parameters.