Yezeng Chen · 陈叶增
NLP · Large Language Models · AI for Science
About me
I am an engineer at Huawei Consumer Business Group (XiaoYi / 小艺), working on large language models and intelligent assistants. I received my M.E. in Computer Science from ShanghaiTech University in 2025.
My research interests span natural language processing, mathematical reasoning with LLMs, and AI for Science — especially AI4Chemistry and AI4Math. I believe AI-driven scientific discovery will reshape how we approach fundamental research.
For collaboration or opportunities, feel free to reach out via email.
Experience
-
Huawei · XiaoYi (小艺) — Algorithm Engineer
2025.08 – Present · Consumer Business Group -
Huawei · XiaoYi (小艺) — Research Intern
2024.06 – 2024.12
Education
-
M.E., Computer Science and Technology
ShanghaiTech University · 2022.09 – 2025.06 -
B.E., Computer Science and Technology
ShanghaiTech University · 2018.09 – 2022.06
Publications
按年份倒序排列,与 Google Scholar 同步。
- Z. Zhang, P. Lv, M. Wan, J. Fang, D. Guo, Y. Chen, Y. Liu, W. Ma, J. Sun, et al. Hot-Swap MarkBoard: An Efficient Black-box Watermarking Approach for Large-scale Model Distribution. ACM Multimedia (MM '25), 2025
- X. Li, Z. Feng, Y. Chen, W. Dai, Z. He, Y. Zhou, S. Jiao. Coeff-KANs: A Paradigm to Address the Electrolyte Field with KANs. arXiv:2407.20265, 2024
- W. Dai, Y. Chen, Z. Dai, Z. Huang, Y. Liu, Y. Pan, B. Song, C. Zhong, X. Li, et al. KALE-LM: Unleash the Power of AI for Science via Knowledge and Logic Enhanced Large Model. arXiv:2409.18695, 2024
- Y. Chen, Z. Chen, Y. Zhou. Brain-Inspired Two-Stage Approach: Enhancing Mathematical Reasoning by Imitating Human Thought Processes. arXiv:2403.00800, 2024
- Z. Chen, Y. Chen, J. Han, Z. Huang, J. Qi, Y. Zhou. An Empirical Study of Data Ability Boundary in LLMs' Math Reasoning. arXiv:2403.00799, 2024
- H. Wu, W. Hui, Y. Chen, W. Wu, K. Tu, Y. Zhou. Conic10K: A Challenging Math Problem Understanding and Reasoning Dataset. Findings of EMNLP 2023
- Y. Yangming, L. Chunxiao, C. Yezeng, D. Zijie, Z. Yi. Knowledge Discovery from Natural Languages: A Linguistic Dataset of 10K Kinship Relations. IEEE BDAI, 2023
Projects
A-Share Quant Dashboard
Python pipeline + static HTML dashboard: market overview, capital flow, multi-factor stock screening, candlestick patterns, and backtesting.
2026 World Cup Predictor
Probability-based group-stage predictions for 48 teams / 72 matches, with walk-forward backtest on 1998–2022 tournaments.
Copper Arbitrage Monitor
Real-time basis, regional premium, import P&L, and cross-term spread monitoring for electrolytic copper spot arbitrage.
More on GitHub
Llama-Factory forks, RAG experiments, evaluation toolkits, and other research repos.