- 🍁 I'm Xuefen (雪枫 in Chinese), a final-year Ph.D. student @Institute of Computing Technology of Chinese Academy of Sciences, received my bachelor degree@BUPT in the summer of 2021.
- 🌱 My research interest mainly includes
weakly supervised learning,misinformation,AutoDriveandvarious deep learning applications. Previously, I've gained industry experience at AMD, Li Auto and Meituan. I'm passionate about uncovering valuable & unsettled research topics, so perhaps we can collaborate to do some interesting things😆. - 📫 You can reach me via my email: jiangxf0903@gmail.com, visit my personal website, or view my Google scholar.
- 🌼 I love films & anime & sightseeing & music & night & sleeping a lot.
- ✨ I'm the big fan of Iris.
💬 Recent News~ (more updates in here)
- 🐱 [2026.4] Our Zofia (about misinformation) is accepted by ACL'26 Findings. Meanwhile, our TriTAN about robust image retrieval is accepted by IEEE TBD.
- 🥳 [2026.2] It is excited to share our twins TransDiffuser&LightVLA are accepted by ICRA'26, see you in Vienna!
- 🧩 [2025.12] Our research paper FedROM got accepted by IEEE TMC. Anyway, 2025.12.18 should be remembered💦.
- 🤖 [2025.9] We propose LightVLA which achieves the state-of-the-art performance on the LIBERO benchmark with significantly reduced inferency overhead.
- 🚗 [2025.5] Our TransDiffuser achieves the new state-of-the-art performance in the NAVSIM leaderboard.
- 📕 [2024.12] We release one experimental study LLM4CVD which exploits LLMs to identify vulnerabilities within codes.
- 🎉 [2024.7] One full research paper got accepted @ACM/CIKM'24. We also release one related comprehensive benchmark study FNBench and another follow-up work. Another work which focuses on long-tailed learning got accepted @ACML'24 oral.
- 🚗 [2024.6] We reach top 5 teams in CVPR 2024 AutoDrive Grand Challenge (Driving with language track) with offline labels and multi-modal LLM. Another work (AMD's SLAM solution for underground autonomous parking) is released here.





