Xinyu Duan

Researcher, Huawei Technologies Ltd.

Joined

July 2019

Names

Xinyu Duan (Preferred)

Emails

****@huawei.com (Confirmed)

Personal Links

Google Scholar

DBLP

Semantic Scholar

Career & Education History

Researcher

Huawei Technologies Ltd. (huawei.com)

2019 – 2026

PhD student

Zhejiang University (zju.edu.cn)

2014 – 2019

Undergrad student

Zhejiang University (zju.edu.cn)

2010 – 2014

Advisors, Relations & Conflicts

PhD Advisor

Fei Wu

2014 – 2019

PhD Advisor

Zhou Zhao

2014 – 2019

Expertise

LLM Inference Optimization

2022 – 2025

LLM

2022 – 2025

Pre-trained Language Model

2019 – 2025

Natural Language Processing

2015 – 2022

Question Answering

2015 – 2022

Text Retrieval

2015 – 2022

Publications

CaliDrop: KV Cache Compression with Query-based Calibration
Yi Su, Quantong Qiu, Yuechi Zhou, Qingrong Xia, Ping Li, Xinyu Duan, Zhefeng Wang, Juntao Li, Min Zhang
- Submitted to ICLR 2026
- Readers: Everyone
Direction-Magnitude Decoupling for Fast Video Generation with Flow Matching Models
Haonan Xu, Feiyang Chen, Chen SongKui, Zhefeng Wang, Xinyu Duan, Zhipeng Liang, Baoxing Huai, Yang Yang
- Submitted to ICLR 2026
- Readers: Everyone
SinkQ: Accurate 2-bit KV Cache Quantization with Dynamic Sink Tracking
Yi Su, Yuechi Zhou, Juntao Li, Qingrong Xia, Ping Li, Xinyu Duan, Zhefeng Wang, Min Zhang
- ICLR 2025 Conference Withdrawn Submission
- Readers: Everyone
Beware of Calibration Data for Pruning Large Language Models
Yixin Ji, Yang Xiang, Juntao Li, Qingrong Xia, Ping Li, Xinyu Duan, Zhefeng Wang, Min Zhang
- ICLR 2025 Poster
- Readers: Everyone
CASD: Enhancing Generation Accuracy via Context-Aware Speculative Decoding
Jikai Wang, Zhenxu Tian, Juntao Li, Qingrong Xia, Xinyu Duan, Zhefeng Wang, Baoxing Huai, Min Zhang
- ICLR 2025 Conference Withdrawn Submission
- Readers: Everyone
When and How to Grow? On Efficient Pre-training via Model Growth
Jikai Wang, Juntao Li, Min Zhang, Zechang Li, Qingrong Xia, Xinyu Duan, Zhefeng Wang, Baoxing Huai
- ACML 2024 Conference Track
- Readers: Everyone
Poisoning for Debiasing: Fair Recognition via Eliminating Bias Uncovered in Data Poisoning
Yi Zhang, Zhefeng Wang, Rui Hu, Xinyu Duan, Yi ZHENG, Baoxing Huai, Jiarun Han, Jitao Sang
- MM2024 Poster
- Readers: Everyone
Revisiting Pre-trained Language Models and their Evaluation for Arabic Natural Language Processing
Abbas Ghaddar, Yimeng Wu, Sunyam Bagga, Ahmad Rashid, Khalil Bibi, Mehdi Rezagholizadeh, Chao Xing, Yasheng Wang, Xinyu Duan, Zhefeng Wang, Baoxing Huai, Xin Jiang, Qun Liu, Philippe Langlais
- 07 Dec 2023
- OpenReview Archive Direct Upload
- Readers: Everyone
Are Bert Family Good Instruction Followers? A Study on Their Potential And Limitations
yisheng xiao, Juntao Li, Zechen Sun, Zechang Li, Qingrong Xia, Xinyu Duan, Zhefeng Wang, Min Zhang
- ICLR 2024 poster
- Readers: Everyone
AuG-KD: Anchor-Based Mixup Generation for Out-of-Domain Knowledge Distillation
Zihao TANG, Zheqi Lv, Shengyu Zhang, Yifan Zhou, Xinyu Duan, Fei Wu, Kun Kuang
- ICLR 2024 poster
- Readers: Everyone

View all 13 publications

Co-Authors

View all 56 co-authors