Xiangpeng Wei (魏相鹏) Algorithm Engineer at ByteDance Seed Location: Hangzhou, China | Email: pemywei@gmail.com [Google Scholar] [DBLP] Feel free and welcome to contact for intern positions and possible collaboration!

Biography

About Me: I am currently an Algorithm Engineer at ByteDance Seed. I received my Ph.D. degree from the University of Chinese Academy of Sciences (UCAS) in 2021, supervised by Prof. Yue Hu.

My research interests include large language models and multimodal applications. Once awarded the Outstanding Paper Award at ACL 2022 and recognized as a Beijing Excellent Graduate.

Education

• 2016.9-2021.6        Ph.D in Computer Science, University of Chinese Academy of Sciences.
• 2012.9-2016.6        B.S. in Software Engineering, Dalian University of Technology.

News

• 2025.04        We are pleased to announce the launch of our new reasoning model, Seed-thinking-v1.5.
• 2023.09        Our paper "EMMA-X: An EM-like Multilingual Pre-training Algorithm for Cross-lingual Representation Learning" has been accepted to NeurIPS 2023 main conference.
• 2023.07        We are pleased to announce the launch of our new polyglot large language model, PolyLM.
• 2022.05        Our paper "Learning to Generalize to More: Continuous Semantic Augmentation for Neural Machine Translation" has been selected as an outstanding paper in ACL 2022.
• 2022.02        Our paper "Learning to Generalize to More: Continuous Semantic Augmentation for Neural Machine Translation" has been accepted to ACL 2022 main conference.

Publications

• Bytedance Seed. "Seed1.5-Thinking: Advancing Superb Reasoning Models with Reinforcement Learning". [paper]
• Bytedance Seed. "DAPO: An Open-Source LLM Reinforcement Learning System at Scale". [paper][page]
• Ping Guo, Xiangpeng Wei*, Yue Hu*, Baosong Yang, Dayiheng Liu, Fei Huang and Jun Xie. "EMMA-X: An EM-like Multilingual Pre-training Algorithm for Cross-lingual Representation Learning". In NeurIPS 2023.
• Xiangpeng Wei, Haoran Wei, Huan Lin, Tianhao Li, Pei Zhang, Xingzhang Ren, Mei Li, Yu Wan, Zhiwei Cao, Binbin Xie, Tianxiang Hu, Shangjie Li, Binyuan Hui, Bowen Yu, Dayiheng Liu, Baosong Yang, Fei Huang, Jun Xie. "PolyLM: An Open Source Polyglot Large Language Model". [paper][huggingface][modelscope]
• Xiangpeng Wei, Heng Yu, Yue Hu, Rongxiang Weng, Weihua Luo, Jun Xie and Rong Jin. "Learning to Generalize to More: Continuous Semantic Augmentation for Neural Machine Translation". In ACL 2022. Outstanding Paper Award. [paper][code][slides]
• Xiangpeng Wei, Yue Hu, Rongxiang Weng, Luxi Xing, Heng Yu and Weihua Luo. "On Learning Universal Representations Across Languages". In ICLR 2021. [paper]
• Xiangpeng Wei, Heng Yu, Yue Hu, Rongxiang Weng, Luxi Xing and Weihua Luo. "Uncertainty-Aware Semantic Augmentation for Neural Machine Translation". In EMNLP 2020. [paper][code]
• Xiangpeng Wei, Heng Yu, Yue Hu, Yue Zhang, Rongxiang Weng and Weihua Luo. "Multiscale Collaborative Deep Models for Neural Machine Translation". In ACL 2020. [paper][code]
• Xiangpeng Wei, Yue Hu, Luxi Xing, Yipeng Wang and Li Gao. "Translating with Bilingual Topic Knowledge for Neural Machine Translation". In AAAI 2019. [paper]
• Xiangpeng Wei, Yue Hu, Luxi Xing and Li Gao. "Unsupervised Neural Machine Translation with Future Rewarding". In CoNLL 2019. [paper]
• Rongxiang Weng, Heng Yu, Xiangpeng Wei and Weihua Luo. "Towards Enhancing Faithfulness for Neural Machine Translation". In EMNLP 2020.

Experience

• 2024.1-Now            Algorithm Engineer, Bytedance Seed.
• 2021.7-2024.1        Algorithm Expert, Alibaba DAMO Academy.
• 2019.2-2021.7        Research Intern, Alibaba DAMO Academy.

Competitions

• WMT20, ranking 2nd on both German-to-French and French-to-German news shared tasks.
• CCMT19, ranking 2nd on the multi-lingual zero-shot MT task among all constrained systems.

Academic Activities

• Conference program committee: ICLR, EMNLP, ICML, ACL, NeurIPS, NAACL
• Talks:

  Learning to Generalize to More: Continuous Semantic Augmentation for Neural Machine Translation. AI TIME. 2022.06.

  On Learning Universal Representations Across Languages. XTREME Team, Google Inc. 2021.04.

  Multiscale Collaborative Deep Models for Neural Machine Translation. DAMO Academy, Alibaba Group. 2020.07.

Awards

• 2022, ACL 2022 Outstanding Paper Award
• 2021, Beijing Outstanding Graduates
• 2021, Zhu Li Yuehua Outstanding Doctoral Scholarship
• 2020, CAS Presidential Scholarship
• 2019, National Scholarship