WS
Weiqiao Shan
cs.CLcs.AIcs.SDeess.AScs.LG
On Valency
published · living versionsW_kruxywyh·v1 · currentpublished
The Right Source Is Not Enough: Understanding History Interference in Speech-LLM Memory
with HongJin Song, Runwu Shi, Yifei Wu, Yujin Wang +9
1 version
W_af3g4cpf·v1published
Preprints & journals
10 papers in the corpus · 2023–2026Can Terminal Agents Trust Their Own Verification? Diagnosing and Improving Self-Verification2609.38812v1 · Yingfeng Luo, Shaowei Wei, Daixin Wang et al.2026 · 0 citationsarXiv
Rethinking Length-Based Training: Batch Composition and Loss Normalization in Speech Token Language Models2609.25890v1 · Hongjin Song, Runwu Shi, Weiqiao Shan et al.2026 · 0 citationsarXiv
Joint Residual Reweighting for Classifier Free Guidance in Flow-Matching Zero-Shot TTS2606.25672v3 · Runwu Shi, Yujin Wang, Hongjin Song et al.2026 · 0 citationsarXiv
SPRI: SVD-Partitioned Residual Initialization for Data-Constrained MoE Upcycling2606.16456v1 · Weiqiao Shan, Ruixiang Mao, Yuang Li et al.2026 · 0 citationsarXiv
One Size Does Not Fit All: Distribution-Guided Tensor-Wise Adaptive Sparsification for Model Merging2508.06163v2 · Yingfeng Luo, Dingyang Lin, Junxin Wang et al.2025 · 0 citationsarXiv
Leveraging Unit Language Guidance to Advance Speech Modeling in Textless Speech-to-Speech Translation2505.15333v1 · Yuhao Zhang, Xiangnan Ma, Kaiqi Kou et al.2025 · 1 citationarXiv
Enhancing Speech Large Language Models with Prompt-Aware Mixture of Audio Encoders2502.15178v3 · Weiqiao Shan, Yuang Li, Yuhao Zhang et al.2025 · 3 citationsarXiv
Optimizing Speech Multi-View Feature Fusion through Conditional Computation2501.08057v1 · Weiqiao Shan, Yuhao Zhang, Yuchen Han et al.2025 · 0 citationsarXiv
Early Exit Is a Natural Capability in Transformer-based Models: An Empirical Study on Early Exit without Joint Optimization2412.01455v1 · Weiqiao Shan, Long Meng, Tong Zheng et al.2024 · 0 citationsarXiv
PartialFormer: Modeling Part Instead of Whole for Machine Translation2310.14921v2 · Tong Zheng, Bei Li, Huiwen Bao et al.2023 · 2 citationsarXiv
Career total: 15 works. 10 are in this corpus.
Profile built from the corpus for this byline.
Author records are still filling in while the Hub is in alpha. If this is your page, you'll be able to claim it soon. Spot a mistake? Tell us.