Zobrazeno 1 - 10
of 80
pro vyhledávání: '"Xu, Jiashu"'
Autor:
Xu, Jiashu
Medical image segmentation plays a crucial role in clinical diagnosis and treatment planning. Although models based on convolutional neural networks (CNNs) and Transformers have achieved remarkable success in medical image segmentation tasks, they st
Externí odkaz:
http://arxiv.org/abs/2410.02523
The advancement of Large Language Models (LLMs) has significantly impacted various domains, including Web search, healthcare, and software development. However, as these models scale, they become more vulnerable to cybersecurity risks, particularly b
Externí odkaz:
http://arxiv.org/abs/2409.19993
The security of multi-turn conversational large language models (LLMs) is understudied despite it being one of the most popular LLM utilization. Specifically, LLMs are vulnerable to data poisoning backdoor attacks, where an adversary manipulates the
Externí odkaz:
http://arxiv.org/abs/2407.04151
Autor:
Ge, Yunhao, Tang, Yihe, Xu, Jiashu, Gokmen, Cem, Li, Chengshu, Ai, Wensi, Martinez, Benjamin Jose, Aydin, Arman, Anvari, Mona, Chakravarthy, Ayush K, Yu, Hong-Xing, Wong, Josiah, Srivastava, Sanjana, Lee, Sharon, Zha, Shengxin, Itti, Laurent, Li, Yunzhu, Martín-Martín, Roberto, Liu, Miao, Zhang, Pengchuan, Zhang, Ruohan, Fei-Fei, Li, Wu, Jiajun
The systematic evaluation and understanding of computer vision models under varying conditions require large amounts of data with comprehensive and customized labels, which real-world vision datasets rarely satisfy. While current synthetic data gener
Externí odkaz:
http://arxiv.org/abs/2405.09546
Autor:
Xu, Jiashu
Automatic medical image segmentation technology has the potential to expedite pathological diagnoses, thereby enhancing the efficiency of patient care. However, medical images often have complex textures and structures, and the models often face the
Externí odkaz:
http://arxiv.org/abs/2405.05007
The exorbitant cost of training Large language models (LLMs) from scratch makes it essential to fingerprint the models to protect intellectual property via ownership authentication and to ensure downstream users and developers comply with their licen
Externí odkaz:
http://arxiv.org/abs/2401.12255
Autor:
Zhao, Brian Nlong, Xiao, Yuhang, Xu, Jiashu, Jiang, Xinyang, Yang, Yifan, Li, Dongsheng, Itti, Laurent, Vineet, Vibhav, Ge, Yunhao
The popularization of Text-to-Image (T2I) diffusion models enables the generation of high-quality images from text descriptions. However, generating diverse customized images with reference visual attributes remains challenging. This work focuses on
Externí odkaz:
http://arxiv.org/abs/2312.14216
Existing studies in backdoor defense have predominantly focused on the training phase, overlooking the critical aspect of testing time defense. This gap becomes particularly pronounced in the context of Large Language Models (LLMs) deployed as Web Se
Externí odkaz:
http://arxiv.org/abs/2311.09763
We propose a new paradigm to automatically generate training data with accurate labels at scale using the text-to-image synthesis frameworks (e.g., DALL-E, Stable Diffusion, etc.). The proposed approach1 decouples training data generation into foregr
Externí odkaz:
http://arxiv.org/abs/2309.05956
We investigate security concerns of the emergent instruction tuning paradigm, that models are trained on crowdsourced datasets with task instructions to achieve superior performance. Our studies demonstrate that an attacker can inject backdoors by is
Externí odkaz:
http://arxiv.org/abs/2305.14710