Model X-ray:Detecting Backdoored Models via Decision Boundary

Kavli Affiliate: Ting Xu | First 5 Authors: Yanghao Su, Jie Zhang, Ting Xu, Tianwei Zhang, Weiming Zhang | Summary: Backdoor attacks pose a significant security vulnerability for deep neural networks (DNNs), enabling them to operate normally on clean inputs but manipulate predictions when specific trigger patterns occur. Currently, post-training backdoor detection approaches often operate […]


Continue.. Model X-ray:Detecting Backdoored Models via Decision Boundary

Multi-Task Contrastive Learning for 8192-Token Bilingual Text Embeddings

Kavli Affiliate: Feng Wang | First 5 Authors: Isabelle Mohr, Markus Krimmel, Saba Sturua, Mohammad Kalim Akram, Andreas Koukounas | Summary: We introduce a novel suite of state-of-the-art bilingual text embedding models that are designed to support English and another target language. These models are capable of processing lengthy text inputs with up to 8192 […]


Continue.. Multi-Task Contrastive Learning for 8192-Token Bilingual Text Embeddings

An Integrated Data Processing Framework for Pretraining Foundation Models

Kavli Affiliate: Feng Wang | First 5 Authors: Yiding Sun, Feng Wang, Yutao Zhu, Wayne Xin Zhao, Jiaxin Mao | Summary: The ability of the foundation models heavily relies on large-scale, diverse, and high-quality pretraining data. In order to improve data quality, researchers and practitioners often have to manually curate datasets from difference sources and […]


Continue.. An Integrated Data Processing Framework for Pretraining Foundation Models

An Integrated Data Processing Framework for Pretraining Foundation Models

Kavli Affiliate: Feng Wang | First 5 Authors: Yiding Sun, Feng Wang, Yutao Zhu, Wayne Xin Zhao, Jiaxin Mao | Summary: The ability of the foundation models heavily relies on large-scale, diverse, and high-quality pretraining data. In order to improve data quality, researchers and practitioners often have to manually curate datasets from difference sources and […]


Continue.. An Integrated Data Processing Framework for Pretraining Foundation Models

Regioselective On-Surface Synthesis of [3]Triangulene Graphene Nanoribbons

Kavli Affiliate: Michael F. Crommie | First 5 Authors: Michael C. Daugherty, Peter H. Jacobse, Jingwei Jiang, Joaquim Jornet-Somoza, Reis Dorit | Summary: The integration of low-energy states into bottom-up engineered graphene nanoribbons (GNRs) is a robust strategy for realizing materials with tailored electronic band structure for nanoelectronics. Low-energy zero-modes (ZMs) can be introduced into […]


Continue.. Regioselective On-Surface Synthesis of [3]Triangulene Graphene Nanoribbons

Distinguishable-particle Glassy Crystal: the simplest molecular model of glass

Kavli Affiliate: Feng Wang | First 5 Authors: Leo S. I. Lam, Gautham Gopinath, Zichen Zhao, Shuling Wang, Chun-Shing Lee | Summary: The nature of glassy dynamics and the glass transition are long-standing problems under active debate. In the presence of a structural disorder widely believed to be an essential characteristic of structural glass, identifying […]


Continue.. Distinguishable-particle Glassy Crystal: the simplest molecular model of glass

Robust-Wide: Robust Watermarking against Instruction-driven Image Editing

Kavli Affiliate: Ting Xu | First 5 Authors: Runyi Hu, Jie Zhang, Ting Xu, Tianwei Zhang, Jiwei Li | Summary: Instruction-driven image editing allows users to quickly edit an image according to text instructions in a forward pass. Nevertheless, malicious users can easily exploit this technique to create fake images, which could cause a crisis […]


Continue.. Robust-Wide: Robust Watermarking against Instruction-driven Image Editing

Robust-Wide: Robust Watermarking against Instruction-driven Image Editing

Kavli Affiliate: Ting Xu | First 5 Authors: Runyi Hu, Jie Zhang, Ting Xu, Jiwei Li, Tianwei Zhang | Summary: Instruction-driven image editing allows users to quickly edit an image according to text instructions in a forward pass. Nevertheless, malicious users can easily exploit this technique to create fake images, which could cause a crisis […]


Continue.. Robust-Wide: Robust Watermarking against Instruction-driven Image Editing

Low-Dose CT Reconstruction Using Dataset-free Learning

Kavli Affiliate: Feng Wang | First 5 Authors: Feng Wang, Renfang Wang, Bo Yang, Hong Qiu, | Summary: Low-Dose computer tomography (LDCT) is an ideal alternative to reduce radiation risk in clinical applications. Although supervised-deep-learning-based reconstruction methods have demonstrated superior performance compared to conventional model-driven reconstruction algorithms, they require collecting massive pairs of low-dose and […]


Continue.. Low-Dose CT Reconstruction Using Dataset-free Learning