3 papers
cs.CV2026
SIEFormer: Spectral-Interpretable and -Enhanced Transformer for Generalized Category Discovery
Chunming Li, Shidong Wang, Tong Xin +1
This paper presents a novel approach, Spectral-Interpretable and -Enhanced Transformer (SIEFormer), which leverages spectral analysis to reinterpret the attention mechanism within…
cs.CL2025
InfiMed: Low-Resource Medical MLLMs with Advancing Understanding and Reasoning
Zeyu Liu, Zhitian Hou, Guanghao Zhu +3
Multimodal Large Language Models (MLLMs) have achieved remarkable progress in domains such as visual understanding and mathematical reasoning. However, their application in the med…
cs.CV2025
Few-shot Novel Category Discovery
Chunming Li, Shidong Wang, Haofeng Zhang
The recently proposed Novel Category Discovery (NCD) adapt paradigm of transductive learning hinders its application in more real-world scenarios. In fact, few labeled data in part…