Xin Jin (金 鑫)
Research Assistant | HangZhou, China
Xin Jin (金 鑫)

About Me 😁

I am a research assistant at Westlake University, supervised by Prof. Huan Wang. I received my M.S. degree in Computer Science from Chongqing Technology and Business University in 2025, under the supervision of Prof. Huafeng Qin. In 2022, I obtained my B.E. degree in Internet of Things Engineering from Pass Collage of Chongqing Technology and Business University.

Research: Currently, my research centers on Efficient AI, Large Language Models, and Multi-modal Large Language Models. Previously, my research focused on Data Augmentation and Biometric Identification.

Interests: In my spare time, I am an amateur photographer with a passion for capturing scenery and starry sky shots 🌌. I use both digital cameras 📷 and film cameras 🎞️ for my photography.

Feel free to contact me for academic and research discussions. Please drop me an email at jinxin86@westlake.edu.cn.

I am actively seeking 26 Fall or 27 Spring PhD opportunities!

News 💬

  • [March, 2026] 📕 [Preprint] One first-author paper on palm vein identification StarLKNet, is released.
  • [March, 2026] 📕 [Preprint] Our co-author paper on omini understanding benchmark LVOmniBench, is released on arXiv.
  • [Feb, 2026] 🎉 [CVPR2026] One co-first author paper on efficient video understanding, EarlyTom is accpted by CVPR 2026, congrats to Hesong Wang.
  • [Jan, 2026] 🎉 One paper on efficient mixup augmentation and preference tuning, MergeMix is accpted by ICLR 2026.
  • [Oct, 2025] 📕 [ICLR2026] One paper on Mixup & MLLM, MergeMix is released on arXiv.
  • [Feb, 2025] 🎉 [ACL2025] One co-author paper on (M)LLM optimizers, SGG is accepted by ACL 2025 main, congrats to Siyuan Li, Juanxi Tian and Zedong Wang.
  • [Feb, 025] 🎉 [IEEE TIM] Two co-author papers on eye moment identification, EMMixFormer and EM-DARTS are accepted by IEEE TIM, congrats to Hongyu Zhu.
  • [Nov, 2024] 🎉 [CCBR2024] One co-first author paper on eye moment identification, Relax DARTS, is accepted by CCBR 2024 (Oral), congrats to Hongyu Zhu.
  • [Jul, 2024] 📕 [Preprint] One survey on mixup data augmentation, Survey, is released on arXiv.
  • [Jul, 2024] 🎉 [ECCV2024] One paper on mixup data augmentation, SUMix, is accepted by ECCV 2024.
  • [May, 2024] 🎉 [HSI2024] One co-first author paper on vein identification, GANet, is accepted by HSI 2024 (Oral), congrats to Hongchao Liao.
  • [Jan, 2024] 🎉 [ICLR2024] One paper on mixup data augmentation, AdAutoMix, is accepted by ICLR 2024 (Spotlight).

Experiences 📝

 

Research Intern | Chongqing Financial Institute
Time: July 2023 - Sept 2023. Advisor: Prof. Huafeng Qin.

 

Research Intern | Westlake University
Time: Mar. 2024 -- July 2025. Advisor: Prof. Stan Z. Li.

 

Research Assistant | Westlake University
Time: July. 2025 -- Present. Advisor: Prof. Huan Wang.

Publications 📖

Selected Preprints:

 

LVOmniBench: Pioneering Long Audio-Video Understanding Evaluation for Omnimodal LLMs
Keda Tao🌟, Yuhua Zhen🌟, Jia Xu, Wenjie Du, Kele Shao, Hesong Wang, Xueyi Chen, Xin Jin, Junhan Zhu, Bohan Yu, Weiqiang Wang, Jian Liu, Can Qin, Yulun Zhang, Ming-Hsuan Yang, and Huan Wang✉️
arXiv, 2026 [arXiv] [PDF] [Dataset] [Webpage]

 

A Survey on Mixup Augmentations and Beyond
Xin Jin🌟, Hongyu Zhu🌟, Siyuan Li🌟, Zedong Wang, Zicheng Liu, Juanxi Tian, Chang Yu, Huafeng Qin, and IEEE Fellow Stan. Z. Li✉️
arXiv, 2024 [arXiv] [Awesome] [BibTeX]

 

Unleashing Palm Vein Recognition via Smooth Mask Mixup and Large Kernel-based Network
Xin Jin🌟, Hongyu Zhu🌟, Mounim A. El-Yacoubi, Haiyang Li, Xingrong Fan, Lin Chen, Yun Jiang✉️, and Qun Song✉️
arXiv, 2024 [PDF] </p></td></tr> </table> **Conferences (🌟: Equal Contribution. ✉️: Corresponding Author.):**

 

EarlyTom: Early Token Compression Completes Fast Video Understanding
Hesong Wang🌟, Xin Jin🌟, Lu Lu, Chenhaowen Li, Jian Chenm Qiang Liu, Huan Wang✉️
CVPR, 2026

 

MergeMix: A Unified Augmentation Paradigm for Visual and Multi-Modal Understanding
Xin Jin🌟, Siyuan Li🌟, Siyong Jian, Kai Yu, Huan Wang✉️
ICLR, 2026 [Website] [arXiv] [HF DailyPapers] [BibTeX]

 

Taming LLMs by Scaling Learning Rates with Gradient Grouping
Siyuan Li🌟, Juanxi Tian🌟, Zedong Wang🌟,Xin Jin, Zicheng Liu✉️, Wentao Zhang, Dan Xu
ACL-main, 2025 [arXiv] [HF DailyPapers Top-4] [BibTeX]

 

SUMix: Mixup with Semantic and Uncertain Information
Huafeng Qin🌟,✉️, Xin Jin🌟, Hongyu Zhu🌟, Hongchao Liao, Mounim A. El-Yacoubi, Xinbo Gao
ECCV, 2024 [arXiv] [Code] [Poster] [BibTeX]

 

Adversarial AutoMixup
Huafeng Qin🌟,✉️, Xin Jin🌟, Yun Jiang, Mounim A. El-Yacoubi, Xinbo Gao
ICLR, 2024 (Spotlight) [arXiv] [Code] [Poster] [BibTeX]

**Journals:**
 

EM-DARTS: Hierarchical Differentiable Architecture Search for Eye Movement Recognition
Huafeng Qin, Hongyu Zhu, Xin Jin, Xin Yu, Mounim A. El-Yacoubi, and Shuqiang Yang
IEEE TIM, 2025 [IEEE] [BibTeX]

 

EmMixformer: Mix Transformer for Eye Movement Recognition
Huafeng Qin, Hongyu Zhu, Xin Jin, Qun Song, Mounim A. El-Yacoubi, and IEEE Fellow Xinbo Gao
IEEE TIM, 2025 [IEEE] [Database] [BibTeX]

### Services 🧸 - **Conference Reviewer:** ICLR (2025-2026), CVPR (2026), ICML (2026), ECCV (2026), ~~NeurIPS (2026)~~ - **Journal Reviewer:** Pattern Recognition. - **Membership** 2024 - 2025 Student Member of IEEE 2024 - 2025 Member of China Computer Federation (CCF) ### Awards 🏆 - Challenge Cup of "Qin Chuang Yuan", National 2nd prize, 2024 - The 14th Challenge Cup, Provincial 2nd prize, 2024 - China International Student Innovation Competition, Provincial 2nd prize, 2024 - The 7th Art Exhibition for College Students in Chongqing, Provincial 1st prize, 2023 - Academic Progress Award, CTBU, 2023-2025 ### Interests 📝 Photography. You could see some photos on 500px website (Long time without update...)