9 papers
CapProbe: Evaluating Detailed Image Captions via Full-Scene Dense Question Answering
Mouxiao Huang, Qiangyu Yan, Borui Jiang +1
Evaluating detailed image captions from Vision-Language Models (VLMs) requires going beyond surface-level semantic similarity. Reference-based metrics (e.g., CIDEr and SPICE) and L…
Focus-then-Context: Subject-Centric Progressive Visual Token Reduction for Vision-Language Models
Yulin Zhao, Zheng Zhang
Vision-Language Models (VLMs) face a bottleneck of prohibitive computational costs arising from massive visual token sequences during inference. Existing vision token reduction met…
Thinking-while-speaking: A Controlled, Interleaved Reasoning Method for Real-Time Speech Generation
Xuan Du, Qiangyu Yan, Wenshuo Li +4
The thinking-while-speaking paradigm aims to make AI communication more human. A key challenge is maintaining fluent speech while performing deep reasoning. Our method, InterRS, ta…
Multimodal Latent Reasoning via Hierarchical Visual Cues Injection
Yiming Zhang, Qiangyu Yan, Borui Jiang +1
The advancement of multimodal large language models (MLLMs) has enabled impressive perception capabilities. However, their reasoning process often remains a "fast thinking" paradig…
PPE: Positional Preservation Embedding for Token Compression in Multimodal Large Language Models
Mouxiao Huang, Borui Jiang, Dehua Zheng +3
Multimodal large language models (MLLMs) have achieved strong performance on vision-language tasks, yet often suffer from inefficiencies due to redundant visual tokens. Existing to…
Towards Lossless Ultimate Vision Token Compression for VLMs
Dehua Zheng, Mouxiao Huang, Borui Jiang +2
Visual language models encounter challenges in computational efficiency and latency, primarily due to the substantial redundancy in the token representations of high-resolution ima…