13 papers
SciFigPlag-Bench: A Benchmark for Provenance-Aware Scientific Figure Plagiarism Detection
Zhiying Cui, Minghao Yang, Linlin Gao +2
Scientific figures often encode the visual evidence behind scientific findings, yet figure plagiarism remains underexplored as a benchmarked multimodal evaluation problem. We prese…
Listen, Do Not Copy: Internalizing Audio-Grounded Scaffold Context for Robust Omni-Model Speech Understanding
Pengfei Zhang, Biao Tian, Tianxin Xie +3
Omni models transcribe clean, single-speaker speech well, but their accuracy drops sharply when speakers overlap and the scene is noisy, exactly where knowing who said what matters…
Branch-JEPA: Finite-Support Predictive Distributions for JEPA World Models
Zhi Song, Ximing Xing, Zhenchao Tang +9
Joint-embedding predictive architectures (JEPAs) learn dynamics by predicting future observations in representation space. Yet most JEPA world models return one latent successor, e…
AG-REPA: Causal Layer Selection for Representation Alignment in Audio Flow Matching
Pengfei Zhang, Tianxin Xie, Minghao Yang +1
REPresentation Alignment (REPA) improves the training of generative flow models by aligning intermediate hidden states with pretrained teacher features, but its effectiveness in to…
L2R: Low-Rank and Lipschitz-Controlled Routing for Mixture-of-Experts
Minghao Yang, Ren Togo, Guang Li +2
Mixture-of-Experts (MoE) models scale neural networks by conditionally activating a small subset of experts, where the router plays a central role in determining expert specializat…
Aligning LLMs with Biomedical Knowledge using Balanced Fine-Tuning
Zhenchao Tang, Fang Wang, Haohuai He +12
Engineering LLMs to accelerate life sciences research requires a robust alignment with biomedical knowledge. We observe that biomedical text exhibits a fundamentally different unce…