Full-duplex speech models execute dense model inference for every audio frame throughout a live interaction. This uniform framewise schedule creates sustained GPU load across dialogue states whose computation has strong, reusable temporal structure. ReFrame formulates full-duplex inference as a silence-aware computatio...
Zhen-Hui Zhou, Guan-Yi Chen, Jin-Tiao Wei et al.· 0 citations
STORM-Bench shows that task accuracy masks gaps in epistemic reliability and state tracking, and introduces STORM-BR, a harmonic metric over joint answer-status correctness that exposes abstention failures masked by aggregate accuracy, alongside STORM-BR-ATTR for uncertainty attribution.
Si-Ru Zhong, Sheng-Han Tan, Rihong Yan et al.· 0 citations
Accurate daily precipitation forecasts are important for weather-sensitive decisions. Although recent Weather Foundation Models (WFMs) have achieved significant progress, they are trained on reanalysis data, which exhibit regional and intensity-dependent biases. Since retraining WFMs on observation-based targets is cos...
Guo-Wen Li, Yang Liu, Meng-Xuan Chen et al.· 0 citations
CompKV is introduced, the first compensation-aware sparse attention framework that divides tokens into blocks and explicitly optimizes selection for the downstream compensation mechanism, and shows that the residual left by block-level mean compensation is governed by both block attention mass and within-block logit va...
Zheng-Hong Huang, Rui-Zhe Yao, Dan-Yi Liu et al.· arXiv.org· 1 citation· ⚡1
Prefix-Steered Recurrent Memory (PREM), a memory-token-free framework for frozen vision-language models (VLMs) that separates video ingestion from query answering, consistently outperforms frozen baselines at every evaluated visual budget.
This work introduces the entropy momentum pseudo-labeling (EMP) to refine k-means assignments by leveraging entropy-aware confidence and temporal prediction momentum, and utilizes the contextual neighborhood topology (CNT) to exploit the intrinsic geometric structure of the target feature space.
Qing-Mei Li, Jue-Peng Zheng, Jia-Rui Zhang et al.· 0 citations
GCR is proposed, a training-free framework that casts fixed-budget frame selection as a joint evidence curation problem and maintains a fixed number of chronologically ordered frames and requires no VLM training or architectural modification.
Fan Wei, Si-Ru Zhong, Run-Min Dong et al.· 2 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.