Short Bio
Dr. Tie-Yan Liu graduated from Tsinghua University, where he obtained his PhD degree in 2003. With a career spanning approximately two decades at Microsoft Research Asia, Dr. Liu has served as Assistant Managing Director and led the AI research area. In 2022, he co-founded Microsoft’s first mission-driven global laboratory – Microsoft Research AI for Science, and is now a Distinguished Scientist leading its Asia team. He has held positions as an adjunct professor at Carnegie Mellon University (CMU), Tsinghua University, Nankai University, Hong Kong University of Science and Technology, University of Science and Technology of China, Huazhong University of Science and Technology, Sun Yat-sen University, and as an honorary professor at the University of Nottingham.
Dr. Liu is renowned for his pioneering work on learning to rank, and he has also made significant contributions in deep learning, reinforcement learning, and AI for Science. He is the author of two academic monographs and has published hundreds of papers in top conferences and journals. His work has been cited over 60,000 times, with an H-index of 94. He has served as the conference chair, program committee chair, or (senior) area chair for many top AI conferences (such as WWW/WebConf, SIGIR, NeurIPS, ICLR, ICML, IJCAI, AAAI, KDD, etc.), and as associate editor for renowned international journals like ACM TOIS, ACM TWEB, IEEE TPAMI. He has received numerous awards, including Best Student Paper, Most Cited Paper, Highest Cited Chinese Scholar, and Most Influential Scholar. The International Open Benchmark Council has recognized him as one of the world’s top 100 AI scholars since 1943. In 2016, he was named an IEEE Fellow, in 2021 an ACM Fellow, and in 2022 an AAIA Fellow.
Recent major research achievements of Dr. Liu’s team include:
- In 2016, they published LightGBM, which has been cited more than 12,000 times since then. It has become the most commonly used AI tool in Kaggle competitions, KDD Cup, and industry decision-making, and was selected by the International Open Benchmark Council into the top 100 AI achievements since 1940s.
- In 2018, they invented dual learning, which achieved human parity in Chinese-English news translation for the first time. The following year, it won eight championships at the WMT Machine Translation Competition and is now the core technology of Microsoft Azure machine translation service.
- In 2019, they developed Mahjong AI Suphx, which achieved the 10 DAN for the first time on the famous Mahjong platform “Tenhou,” significantly surpassing top human players and causing a significant stir in the Mahjong community.
- In 2021, they invented Graphormer for molecular modeling, which won championships in the inaugural OGB-LSC molecular modeling challenge and OC20 Open Catalyst Challenge.
- In 2022, they released BioGPT for understanding biomedical literature, which reached human parity on the PubMed QA task for the first time.
- In 2023, they developed TamGen for drug design, which, in close collaboration with Global Health Drug Discovery Institute (GHDDI) and Bill & Melinda Gates Foundation, designed novel and highly effective candidate drugs for tuberculosis and coronaviruses. Following laboratory synthesis and enzyme inhibition tests, the biological activity of these candidate drugs was found to be nearly 10 times higher than known lead compounds.
- In 2024, they published Distributional Graphormer, which achieved end-to-end prediction of the equilibrium distribution of molecular structures, establishing a bridge between the microscopic structures of molecules and their macroscopic chemical and biological properties.
———
刘铁岩博士,1994-2003年就读于清华大学,并获得博士学位。毕业后加入微软亚洲研究院,曾任副院长,领导其AI研究团队;2022年联合创立了微软公司首个使命驱动的全球实验室 – 微软研究院科学智能中心,并担任杰出首席科学家,领导其亚洲研究团队。他(曾)被聘为卡内基梅隆大学(CMU)、清华大学、南开大学、香港科技大学、中国科技大学、华中科技大学、中山大学的兼职教授/博士生导师,诺丁汉大学的荣誉教授。
刘博士的先锋性研究促进了机器学习与信息检索的融合,被公认为“排序学习”领域的代表人物。他在深度学习、强化学习、以及人工智能驱动的科学发现等方面也颇有建树,著有两部学术专著、在顶级国际会议和期刊上发表了数百篇论文,包含多篇自然杂志(Nature)及其子刊的论文,被引用六万余次,H-index为94。他曾担任诸多人工智能顶级会议(如WWW/WebConf、SIGIR、NeurIPS、ICLR、ICML、IJCAI、AAAI、KDD等)的大会主席、程序委员会主席、或(资深)领域主席,以及ACM TOIS、ACM TWEB、IEEE TPAMI等知名国际期刊副主编。他曾多次获得最佳论文奖、最高引用论文奖、最高引中国学者奖、全球最有影响力学者奖。他被国际开放测评委员会评为自1943年以来全球最重要的百位人工智能学者之一,于2016年被评为国际电气电子工程师学会会士(IEEE Fellow)、2021年被评为国际计算机学会会士(ACM Fellow)、2022年被评为亚太人工智能学会会士(AAIA Fellow)。
刘博士团队近年来的主要学术成果包括:
- 2016年发表了用于实现复杂决策任务的AI模型LightGBM,其单篇论文引用数超过一万两千次,目前已成为Kaggle比赛、KDD Cup和产业决策中最常用的AI工具,并被国际开放测评委员会评为AI领域发展至今最重要的百项研究成果之一。
- 2018年发明了用于无监督学习的对偶学习技术,在中英新闻翻译任务上首次达到了人类专家水平,并于次年获得WMT国际机器翻译比赛的8项冠军,目前是微软Azure AI机器翻译服务的核心技术。
- 2019年研发了麻将AI Suphx,在国际知名竞技麻将平台“天凤”上首次荣升十段,稳定段位显著超越人类顶级选手,在竞技麻将界引起很大反响,有多部著作专门讨论和分析Suphx的战略战术。
- 2021年发布了用于分子建模的AI算法Graphormer,在首届 OGB-LSC 分子建模国际比赛和 OC20 催化剂设计国际开放挑战赛中获得冠军。
- 2022年发布了用于理解生物医学文献的AI模型BioGPT,在PubMed问答任务上首次达到人类专家水平。
- 2023年开发了用于药物设计的AI模型TamGen,与全球健康药物研发中心(GHDDI)和盖茨基金会深入合作,为肺结核和冠状病毒等肆虐全球的传染病设计出全新的高效候选药物,经过实验室合成和酶抑制试验,与已知的先导化合物相比,生物活性提高了近10倍,在业界引起了很大反响。
- 2024年发表了Distributional Graphormer模型,对分子结构的平衡分布实现了端到端预测,在分子微观结构和宏观物化属性之间建立了桥梁。