Shaolei Zhang (张绍磊) is currently working toward his Ph.D. degree in the Key Laboratory of Intelligent Information Processing, Institute of Computing Technology, Chinese Academy of Sciences (中国科学院计算技术研究所), advised by Yang Feng (冯洋). He received his bachelor’s degree from Beijing University of Posts and Telecommunications in 2020, majoring in computer science and technology (北京邮电大学计算机科学与技术实验班).
His research interests include nature language processing, machine translation, simultaneous translation and large language model. He has published over 10 papers at the top international AI/NLP conferences such as ACL, EMNLP, NeurIPS, ICLR, AAAI. He won the first place in the streaming transcription track of AutoSimTrans 2021.
🔥 News
- 2023.12: 🎉 One paper is accepted by ICASSP 2024!
- 2023.10: 🎉 Two papers are accepted by EMNLP 2023!
- 2023.09: 👏 Serve as Area Chair of ACL/EACL/NAACL ARR 2023!
- 2023.09: 🎉 One paper is accepted by NeurIPS 2023!
- 2023.06: 🎉 Our cross-lingual aligned LLM BayLing is released.
- 2023.05: 🎉 Two papers are accepted by ACL 2023.
- 2023.01: 🎉 One paper is accepted by ICLR 2023 (spotlight)!
- 2022.10: 🎉 Three papers are accepted by EMNLP 2022!
- 2022.02: 🎉 Three papers are accepted by ACL 2022!
📝 Publications
Demo
BayLing: Bridging Cross-lingual Alignment and Instruction Following through Interactive Translation for Large Language Models
Shaolei Zhang, Qingkai Fang, Zhuocheng Zhang, Zhengrui Ma, Yan Zhou, Langlin Huang, Mengyu Bu, Shangtong Gui, Yunji Chen, Xilin Chen, Yang Feng
- BayLing (百聆) is a LLM equipped with advanced language alignment.
- BayLing is the first research to use interactive translation tasks to complete the cross-language transfer of LLM’s generation capabilities.
- BayLing is selected for inclusion in the 2022-2023 Top 100 Opensource achievements: Open100 (2022-2023), launched by the International Open Benchmark Council (BenchCouncil).
Demo
TruthX: Alleviating Hallucinations by Editing Large Language Models in Truthful Space
Shaolei Zhang *, Tian Yu *, Yang Feng
- TruthX is an inference-time method to elicit the truthfulness of LLMs by editing their internal representations in truthful space, thereby mitigating the hallucinations of LLMs.
- TruthX can control LLMs to generate truthful or hallucinatory responses by editing only a vector in truthful space.
- On TruthfulQA benchmark, TruthX yields an average enhancement of 20% in truthfulness across 13 advanced LLMs.
Awesome Simultaneous Translation
Shaolei Zhang
- A repository that collects the tookits, common datasets and paper list related to the research on Simultaneous Translation, including text-to-text machine translation and speech-to-text translation.
2024
- Shaolei Zhang *, Tian Yu *, Yang Feng. TruthX: Alleviating Hallucinations by Editing Large Language Models in Truthful Space.
- Shoutao Guo, Shaolei Zhang, Yang Feng. SiLLM: Large Language Models for Simultaneous Machine Translation.
- Shoutao Guo, Shaolei Zhang, Yang Feng. Glancing Future for Simultaneous Machine Translation.(Oral)
2023
- Shaolei Zhang, Yang Feng. Unified Segment-to-Segment Framework for Simultaneous Sequence Generation.
- Zhengrui Ma, Shaolei Zhang, Shoutao Guo, Chenze Shao, Min Zhang, Yang Feng. Non-autoregressive Streaming Transformer for Simultaneous Translation.(Oral)
- Shoutao Guo, Shaolei Zhang, Yang Feng. Simultaneous Machine Translation with Tailored Reference.
- Shaolei Zhang, Yang Feng. End-to-End Simultaneous Speech Translation with Differentiable Segmentation.
- Shoutao Guo, Shaolei Zhang, Yang Feng. Learning Optimal Policy for Simultaneous Machine Translation via Binary Search.
- Shaolei Zhang, Yang Feng. Hidden Markov Transformer for Simultaneous Machine Translation.(Spotlight)
- Shaolei Zhang, Qingkai Fang, Zhuocheng Zhang, Zhengrui Ma, Yan Zhou, Langlin Huang, Mengyu Bu, Shangtong Gui, Yunji Chen, Xilin Chen, Yang Feng. BayLing: Bridging Cross-lingual Alignment and Instruction Following through Interactive Translation for Large Language Models.
2022
- Shaolei Zhang, Yang Feng. Information-Transport-based Policy for Simultaneous Translation.(Oral)
- Shaolei Zhang, Shoutao Guo, Yang Feng. Wait-info Policy: Balancing Source and Target at Information Level for Simultaneous Machine Translation.
- Shoutao Guo, Shaolei Zhang, Yang Feng. Turning Fixed to Adaptive: Integrating Post-Evaluation into Simultaneous Machine Translation.
- Shaolei Zhang, Yang Feng. Modeling Dual Read/Write Paths for Simultaneous Machine Translation.
- Shaolei Zhang, Yang Feng. Reducing Position Bias in Simultaneous Machine Translation with Length-Aware Framework.
- Shaolei Zhang, Yang Feng. Gaussian Multi-head Attention for Simultaneous Machine Translation.
2021
- Shaolei Zhang, Yang Feng. Universal Simultaneous Machine Translation with Mixture-of-Experts Wait-k Policy.(Oral)
- Shaolei Zhang, Yang Feng. Modeling Concentrated Cross-Attention for Neural Machine Translation with Gaussian Mixture Model.
- Shaolei Zhang, Yang Feng, Liangyou Li. Future-Guided Incremental Transformer for Simultaneous Translation.(Oral)
- Shaolei Zhang, Yang Feng. ICT’s System for AutoSimTrans 2021: Robust Char-Level Simultaneous Translation.(Oral)
2019
- Shaolei Zhang, Gang Lu, Kai Shuang. Opinion Knowledge Injection Network for Aspect Extraction.
🏆 Honors and Awards
- [2022] ICT’s Special Scholarship (Xia Peisu Award) (计算所所长特别奖(夏培肃奖), highest award in ICT/CAS)
- [2022] National Scholarship (国家奖学金)
- [2021] First place in the streaming transcription track of AutoSimTrans 2021
- [2020] Beijing Outstanding Graduates Award (北京市优秀毕业生)
- [2018] Beijing Merit Student (北京市三好学生)
- [2017] National Scholarship (国家奖学金)
👏 Services
- Area Chair of ACL/EACL/NAACL ARR 2023
- Reviewer of ACL/EMNLP/COLING/NAACL/EACL
- Session Chair of Student Seminar in CCL 2024
- Session Chair of Student Seminar in YSSNLP 2024
- 中国中文信息学会青年工作委员会 学生执委会主任
- Programming Chair of CSSNLP 2020/2021/2023
📖 Educations
- 2020.06 - now: Ph.D. Candidate. Nature language processing. Key Laboratory of Intelligent Information Processing, Institute of Computing Technology, Chinese Academy of Sciences.
- 2016.09 - 2020.06: Bachelor’s degree. Computer science and technology. Beijing University of Posts and Telecommunications.
💬 Invited Talks
- “大模型时代的科研选题和实践分享” on MLNLP Academic Seminar [Slides]
- “跨语言对齐增强大模型——百聆” on AI TIME 大模型嘉年华 [Slides] [Video]
- “如何在大模型时代找到科研切入点?” on CCMT 2023 [Slides] [Video]
- “从机器翻译到同声传译:挑战与进展” on MLNLP Academic Seminar [Slides] [Video]
- AI Time Youth Talk for ICLR 2023 [Video]
💻 Internships
- 2019.12 - 2021.12, Huawei Noah’s Ark Lab, industry-university-research collaboration project, China.