A5102726629- Animal Virus Infections Studies
- Hate Speech and Cyberbullying Detection
- Topic Modeling
- Adversarial Robustness in Machine Learning
- Virology and Viral Diseases
- Bacteriophages and microbial interactions
- Microbial infections and disease research
- Generative Adversarial Networks and Image Synthesis
- Animal Disease Management and Epidemiology
- Animal Nutrition and Physiology
- Text Readability and Simplification
- Social Media and Politics
- Plant biochemistry and biosynthesis
- Explainable Artificial Intelligence (XAI)
- China's Ethnic Minorities and Relations
- Salmonella and Campylobacter epidemiology
- Aquaculture disease management and microbiota
- Natural Language Processing Techniques
- Artificial Intelligence in Healthcare and Education
- Viral gastroenteritis research and epidemiology
- Digital Media Forensic Detection
- Artificial Intelligence in Law
- Chromosomal and Genetic Variations
- Epigenetics and DNA Methylation
- Health disparities and outcomes
Yangzhou University
2020-2025
Poultry Research Institute
2013-2025
Helmholtz Center for Information Security
2022-2024
Chinese Academy of Agricultural Sciences
2011-2024
Guangxi University
2023
State Key Laboratory For Conservation and Utilization of Subtropical Agro-Bioresources
2023
The University of Sydney
2023
Beijing Wuzi University
2023
Alibaba Group (United States)
2021
Indiana University Bloomington
2021
The way users acquire information is undergoing a paradigm shift with the advent of ChatGPT. Unlike conventional search engines, ChatGPT retrieves knowledge from model itself and generates answers for users. ChatGPT's impressive question-answering (QA) capability has attracted more than 100 million within short period time but also raised concerns regarding its reliability. In this paper, we perform first large-scale measurement reliability in generic QA scenario carefully curated set 5,695...
The misuse of large language models (LLMs) has drawn significant attention from the general public and LLM vendors. One particular type adversarial prompt, known as jailbreak emerged main attack vector to bypass safeguards elicit harmful content LLMs. In this paper, employing our new framework JailbreakHub, we conduct a comprehensive analysis 1,405 prompts spanning December 2022 2023. We identify 131 communities discover unique characteristics their major strategies, such prompt injection...
The misuse of large language models (LLMs) has garnered significant attention from the general public and LLM vendors. In response, efforts have been made to align LLMs with human values intent use. However, a particular type adversarial prompts, known as jailbreak prompt, emerged continuously evolved bypass safeguards elicit harmful content LLMs. this paper, we conduct first measurement study on prompts in wild, 6,387 collected four platforms over six months. Leveraging natural processing...
Nowadays, powerful large language models (LLMs) such as ChatGPT have demonstrated revolutionary power in a variety of tasks. Consequently, the detection machine-generated texts (MGTs) is becoming increasingly crucial LLMs become more advanced and prevalent. These ability to generate human-like language, making it challenging discern whether text authored by human or machine. This raises concerns regarding authenticity, accountability, potential bias. However, existing methods for detecting...
State-of-the-art Text-to-Image models like Stable Diffusion and DALLE\cdot2 are revolutionizing how people generate visual content. At the same time, society has serious concerns about adversaries can exploit such to problematic or unsafe images. In this work, we focus on demystifying generation of images hateful memes from models. We first construct a typology consisting five categories (sexually explicit, violent, disturbing, hateful, political). Then, assess proportion generated by four...
Misuse of the Large Language Models (LLMs) has raised widespread concern. To address this issue, safeguards have been taken to ensure that LLMs align with social ethics. However, recent findings revealed an unsettling vulnerability bypassing LLMs, known as jailbreak attacks. By applying techniques, such employing role-playing scenarios, adversarial examples, or subtle subversion safety objectives a prompt, can produce inappropriate even harmful response. While researchers studied several...
Nowadays, powerful large language models (LLMs) such as ChatGPT have demonstrated revolutionary power in a variety of natural processing (NLP) tasks text classification, sentiment analysis, translation, and question-answering. Consequently, the detection machine-generated texts (MGTs) is becoming increasingly crucial LLMs become more advanced prevalent. These ability to generate human-like language, making it challenging discern whether authored by human or machine. This raises concerns...
This study examined whether intense personal feelings toward celebrities are positively correlated with women's consideration of cosmetic surgery, and body dissatisfaction shame serve as possible mediators this relationship. A sample 605 Chinese female undergraduates completed questionnaires on celebrity worship, dissatisfaction, shame. The results revealed a significant positive correlation between the surgery. Moreover, association was mediated both independently by sequentially These...
Large Language Models (LLMs) have raised increasing concerns about their misuse in generating hate speech. Among all the efforts to address this issue, speech detectors play a crucial role. However, effectiveness of different against LLM-generated remains largely unknown. In paper, we propose HateBench, framework for benchmarking on We first construct dataset 7,838 samples generated by six widely-used LLMs covering 34 identity groups, with meticulous annotations three labelers. then assess...
We aimed to identify the characteristic functional components of pigeon eggs and differences among pigeon, chicken, quail eggs. analyzed metabolite profiles three kinds using an untargeted metabolomics-based approach better understand in metabolites Then, we quantitatively validated abundance partial through a targeted approach. A total 692 were identified types 263 significantly differentially abundant found between chicken eggs, The that more than other mainly lipids, lipid-like molecules,...
Newcastle disease (ND) caused by virulent virus (NDV) is an acute, highly contagious and fatal viral affecting most species of birds. Ducks are generally considered to be natural reservoirs or carriers NDV while being resistant strains, even those for chickens; however, ND cases in ducks have been gradually increasing recent years. In the present study, different breeds ages were experimentally infected with duck origin strain duck/Jiangsu/JSD0812/2008 (JSD0812) various routes investigate...
To evaluate the germplasm characteristics and nutritional value of Chinese native pigeon varieties, this study analyzed nutrient composition meat four varieties then compared them with those White King pigeon, which is most commonly used in China. A total 150 pigeons aged 28 d (squabs) 5 breeds including Taihu Shiqi Ta-rim Boot were selected for slaughter. The basic quality parameters contents conventional compositions, inosine acid, amino acids, fatty acids measured. results showed that...
Abstract Cattle and the draught force provided by its skeletal muscle have been integral to agro‐ecosystems of agricultural civilization for millennia. However, relatively little is known about cattle functional genomics (including protein coding genes, non‐coding RNA, etc.). Circular RNAs (circRNAs), as a new class RNAs, can be effectively translated into detectable peptides, which enlightened us on importance circRNAs in physiology function. Here, RNA‐seq, Ribosome profiling (Ribo‐seq),...
Experimental infections of Newcastle disease virus (NDV) strains different avian origin and virulence in mallard (Anas platyrhynchos) ducklings were undertaken to evaluate infectivity pathogenicity NDV for ducks the potential role epidemiology (ND). Ducklings experimentally infected with seven strains, their clinical sign, weight gain, antibody response, shedding, distribution tissues investigated. The duck virulent strain duck/Jiangsu/JSD0812/2008 (JSD0812) Chinese standard F48E8 highly...
Sinophobia, anti-Chinese sentiment, has existed on the Web for a long time. The outbreak of COVID-19 and extended quarantine further amplified it. However, we lack quantitative understanding cause Sinophobia as well how it evolves over In this paper, conduct largescale longitudinal measurement between 2016 2021, two mainstream fringe communities. By analyzing 8B posts from Reddit 206M 4chan’s /pol/, investigate origins, evolution, content Sinophobia. We find that, may be evoked by political...
Text-to-Image generation models have revolutionized the artwork design process and enabled anyone to create high-quality images by entering text descriptions called prompts. Creating a prompt that consists of subject several modifiers can be time-consuming costly. In consequence, trend trading prompts on specialized marketplaces has emerged. this paper, we propose novel attack, namely stealing which aims steal from generated text-to-image models. Successful attacks direct violate...
The naturally isolated avian coronavirus infectious bronchitis virus (IBV) generally cannot replicate in chicken kidney (CK) cells. To explore the molecular mechanism of IBV adapting to CK cells, a series recombinant viruses were constructed by chimerizing S genes cell-adapted strain H120 and non-adapted IBYZ. results showed that S2 subunit determines difference cell tropism two strains. After comparing amino acid sequences protein YZ120, with its parental IBYZ, three substitutions, A138V,...
Recently, the concept of artificial assistants has evolved from science fiction into real-world applications. GPT-4o, newest multimodal large language model (MLLM) across audio, vision, and text, further blurred line between reality by enabling more natural human-computer interactions. However, advent GPT-4o's voice mode may also introduce a new attack surface. In this paper, we present first systematic measurement jailbreak attacks against GPT-4o. We show that GPT-4o demonstrates good...
Image safety classifiers play an important role in identifying and mitigating the spread of unsafe images online (e.g., including violence, hateful rhetoric, etc.). At same time, with advent text-to-image models increasing concerns about AI models, developers are increasingly relying on image to safeguard their models. Yet, performance current remains unknown for real-world AI-generated images. To bridge this research gap, work, we propose UnsafeBench, a benchmarking framework that evaluates...