Ran He
YOU?
Author Swipe
View article: Effects of arbuscular mycorrhizal fungi on biomass and disease-resistance enzyme activities of actinidia arguta against canker
Effects of arbuscular mycorrhizal fungi on biomass and disease-resistance enzyme activities of actinidia arguta against canker Open
In this study, the seedlings of Actinidia arguta were used as the experimental materials. Arbuscular mycorrhizal fungi (AMF), Rhizophagus irregularis, and the canker pathogen Dothiorella gregaria were inoculated. The effects of AMF on the …
View article: Editorial: Motion tracking and deformation analysis in biomechanics
Editorial: Motion tracking and deformation analysis in biomechanics Open
View article: Identification of RNA binding protein genes associated with colorectal cancer by bioinformatics analysis
Identification of RNA binding protein genes associated with colorectal cancer by bioinformatics analysis Open
View article: Test-Time Immunization: A Universal Defense Framework Against Jailbreaks for (Multimodal) Large Language Models
Test-Time Immunization: A Universal Defense Framework Against Jailbreaks for (Multimodal) Large Language Models Open
While (multimodal) large language models (LLMs) have attracted widespread attention due to their exceptional capabilities, they remain vulnerable to jailbreak attacks. Various defense methods are proposed to defend against jailbreak attack…
View article: T^2Agent A Tool-augmented Multimodal Misinformation Detection Agent with Monte Carlo Tree Search
T^2Agent A Tool-augmented Multimodal Misinformation Detection Agent with Monte Carlo Tree Search Open
Real-world multimodal misinformation often arises from mixed forgery sources, requiring dynamic reasoning and adaptive verification. However, existing methods mainly rely on static pipelines and limited tool usage, limiting their ability t…
View article: Survey on generative visual media detection and security
Survey on generative visual media detection and security Open
View article: Leveraging transcriptome-wide association studies identifies the relationship between upper respiratory flora and cell type-specific gene expression in severe respiratory disease
Leveraging transcriptome-wide association studies identifies the relationship between upper respiratory flora and cell type-specific gene expression in severe respiratory disease Open
Objectives The upper respiratory tract flora may influence host immunity and modulate susceptibility to viral respiratory infections. This study aimed to investigate the associations between upper respiratory tract flora and immune cells i…
View article: R-TPT: Improving Adversarial Robustness of Vision-Language Models through Test-Time Prompt Tuning
R-TPT: Improving Adversarial Robustness of Vision-Language Models through Test-Time Prompt Tuning Open
Vision-language models (VLMs), such as CLIP, have gained significant popularity as foundation models, with numerous fine-tuning methods developed to enhance performance on downstream tasks. However, due to their inherent vulnerability and …
View article: Do We Really Need Curated Malicious Data for Safety Alignment in Multi-modal Large Language Models?
Do We Really Need Curated Malicious Data for Safety Alignment in Multi-modal Large Language Models? Open
Multi-modal large language models (MLLMs) have made significant progress, yet their safety alignment remains limited. Typically, current open-source MLLMs rely on the alignment inherited from their language module to avoid harmful generati…
View article: Exploring Vacant Classes in Label-Skewed Federated Learning
Exploring Vacant Classes in Label-Skewed Federated Learning Open
Label skews, characterized by disparities in local label distribution across clients, pose a significant challenge in federated learning. As minority classes suffer from worse accuracy due to overfitting on local imbalanced data, prior met…
View article: Protecting Model Adaptation from Trojans in the Unlabeled Data
Protecting Model Adaptation from Trojans in the Unlabeled Data Open
Model adaptation tackles the distribution shift problem with a pre-trained model instead of raw data, which has become a popular paradigm due to its great privacy protection. Existing methods always assume adapting to a clean target domain…
View article: A Method for Identifying and Assessing Operational Risk Factors of Road Freight E-Commerce Platforms with Multi-Dimensional and Multi-Level Characteristics
A Method for Identifying and Assessing Operational Risk Factors of Road Freight E-Commerce Platforms with Multi-Dimensional and Multi-Level Characteristics Open
Road freight e-commerce platforms, as a specialized form of e-commerce in the road transportation sector, face complex operational risks due to their unique service positioning and business models. This study employs a comprehensive method…
View article: ID-Cloak: Crafting Identity-Specific Cloaks Against Personalized Text-to-Image Generation
ID-Cloak: Crafting Identity-Specific Cloaks Against Personalized Text-to-Image Generation Open
Personalized text-to-image models allow users to generate images of new concepts from several reference photos, thereby leading to critical concerns regarding civil privacy. Although several anti-personalization techniques have been develo…
View article: Appendicitis burden, trends, and inequalities in Europe, 1990–2019: a population-based study
Appendicitis burden, trends, and inequalities in Europe, 1990–2019: a population-based study Open
Background Appendicitis imposes a substantial healthcare burden globally, yet comprehensive insights into its disease burden in Europe remain limited. This study assesses regional trends, disparities and high-burden countries for appendici…
View article: Towards Compatible Fine-tuning for Vision-Language Model Updates
Towards Compatible Fine-tuning for Vision-Language Model Updates Open
So far, efficient fine-tuning has become a popular strategy for enhancing the capabilities of foundation models on downstream tasks by learning plug-and-play modules. However, existing methods overlook a crucial issue: if the underlying fo…
View article: Prototypical Distillation and Debiased Tuning for Black-box Unsupervised Domain Adaptation
Prototypical Distillation and Debiased Tuning for Black-box Unsupervised Domain Adaptation Open
Unsupervised domain adaptation aims to transfer knowledge from a related, label-rich source domain to an unlabeled target domain, thereby circumventing the high costs associated with manual annotation. Recently, there has been growing inte…
View article: How to possess an electronic bill of lading as information? A comparative perspective of the legislation on the “possession problem” of electronic bills of lading
How to possess an electronic bill of lading as information? A comparative perspective of the legislation on the “possession problem” of electronic bills of lading Open
The possession of the paper B/Ls is the basis of the function of a B/L as a document of title in common law systems and the delivery effect of B/Ls in civil law systems. The “possession problem” of eB/Ls is how to ensure that eB/Ls, as int…
View article: Jailbreak Attacks and Defenses against Multimodal Generative Models: A Survey
Jailbreak Attacks and Defenses against Multimodal Generative Models: A Survey Open
The rapid evolution of multimodal foundation models has led to significant advancements in cross-modal understanding and generation across diverse modalities, including text, images, audio, and video. However, these models remain susceptib…
View article: Breaking the Low-Rank Dilemma of Linear Attention
Breaking the Low-Rank Dilemma of Linear Attention Open
The Softmax attention mechanism in Transformer models is notoriously computationally expensive, particularly due to its quadratic complexity, posing significant challenges in vision applications. In contrast, linear attention provides a fa…
View article: ZePo: Zero-Shot Portrait Stylization with Faster Sampling
ZePo: Zero-Shot Portrait Stylization with Faster Sampling Open
Diffusion-based text-to-image generation models have significantly advanced the field of art content synthesis. However, current portrait stylization methods generally require either model fine-tuning based on examples or the employment of…
View article: LoRA-IR: Taming Low-Rank Experts for Efficient All-in-One Image Restoration
LoRA-IR: Taming Low-Rank Experts for Efficient All-in-One Image Restoration Open
Prompt-based all-in-one image restoration (IR) frameworks have achieved remarkable performance by incorporating degradation-specific information into prompt modules. Nevertheless, handling the complex and diverse degradations encountered i…
View article: STAMP: Outlier-Aware Test-Time Adaptation with Stable Memory Replay
STAMP: Outlier-Aware Test-Time Adaptation with Stable Memory Replay Open
Test-time adaptation (TTA) aims to address the distribution shift between the training and test data with only unlabeled data at test time. Existing TTA methods often focus on improving recognition performance specifically for test data as…
View article: An Underwater Image Enhancement Method Based on Diffusion Model Using Dual-Layer Attention Mechanism
An Underwater Image Enhancement Method Based on Diffusion Model Using Dual-Layer Attention Mechanism Open
Diffusion models have been increasingly utilized in various image-processing tasks, such as segmentation, denoising, and enhancement. These models also show exceptional performance in enhancing underwater images. However, conventional mode…
View article: Visual Anchors Are Strong Information Aggregators For Multimodal Large Language Model
Visual Anchors Are Strong Information Aggregators For Multimodal Large Language Model Open
In the realm of Multimodal Large Language Models (MLLMs), vision-language connector plays a crucial role to link the pre-trained vision encoders with Large Language Models (LLMs). Despite its importance, the vision-language connector has b…
View article: Semantic Equitable Clustering: A Simple and Effective Strategy for Clustering Vision Tokens
Semantic Equitable Clustering: A Simple and Effective Strategy for Clustering Vision Tokens Open
The Vision Transformer (ViT) has gained prominence for its superior relational modeling prowess. However, its global attention mechanism's quadratic complexity poses substantial computational burdens. A common remedy spatially groups token…
View article: Band-Attention Modulated RetNet for Face Forgery Detection
Band-Attention Modulated RetNet for Face Forgery Detection Open
The transformer networks are extensively utilized in face forgery detection due to their scalability across large datasets.Despite their success, transformers face challenges in balancing the capture of global context, which is crucial for…
View article: Heterogeneous Test-Time Training for Multi-Modal Person Re-identification
Heterogeneous Test-Time Training for Multi-Modal Person Re-identification Open
Multi-modal person re-identification (ReID) seeks to mitigate challenging lighting conditions by incorporating diverse modalities. Most existing multi-modal ReID methods concentrate on leveraging complementary multi-modal information via f…
View article: DiffMAC: Diffusion Manifold Hallucination Correction for High Generalization Blind Face Restoration
DiffMAC: Diffusion Manifold Hallucination Correction for High Generalization Blind Face Restoration Open
Blind face restoration (BFR) is a highly challenging problem due to the uncertainty of degradation patterns. Current methods have low generalization across photorealistic and heterogeneous domains. In this paper, we propose a Diffusion-Inf…
View article: CSCNET: Class-Specified Cascaded Network for Compositional Zero-Shot Learning
CSCNET: Class-Specified Cascaded Network for Compositional Zero-Shot Learning Open
Attribute and object (A-O) disentanglement is a fundamental and critical problem for Compositional Zero-shot Learning (CZSL), whose aim is to recognize novel A-O compositions based on foregone knowledge. Existing methods based on disentang…
View article: Connecting the Dots: Collaborative Fine-tuning for Black-Box Vision-Language Models
Connecting the Dots: Collaborative Fine-tuning for Black-Box Vision-Language Models Open
With the emergence of pretrained vision-language models (VLMs), considerable efforts have been devoted to fine-tuning them for downstream tasks. Despite the progress made in designing efficient fine-tuning methods, such methods require acc…