俞声，博士，清华大学统计学研究中心副教授，清华大学数据科学研究院 RONG 教授，长期从事医学自然语言处理技术与电子病历分析技术研究。俞声独立开发的电子病历自然语言处理系统被美国哈佛医学院、麻省总医院、退伍军人医学中心等顶尖医学研究机构使用，至今已分析电子病历数十亿篇次。俞声发明的高通量表型提取技术使 i2b2 疾病表型识别算法开发速度从每年1-2个提高到每年超过1000个，并应用于 Veteran Affairs “Million Veteran Program” 等美国国家级精准医学研究项目；该系列论文获评医学信息学顶刊 Journal of the American Medical Informatics Association 的编辑选择奖、国际医学信息学学会2019年年鉴最佳论文奖，并按标准化生物医学实验方法发表于 Nature Protocols。归国后，俞声带领团队围绕中文电子病历和智能诊疗发展了高通量知识图谱构建、无监督中文医学术语发现、医学机器翻译等一系列技术。
Dr. Yu Sheng is an associate professor of Tsinghua University, Center for Statistical Science, and RONG Professor of the Institute of Data Science. His research focuses medical natural language processing (NLP) and data analysis for electronic health records (EHR). The EHR NLP system developed by Dr. Yu has been adopted by top ranked research institutes including Harvard Medical School, Massachusetts General Hospital, and Veteran Affairs Medical Center, to process billions of medical notes. The high-throughput phenotyping technologies developed by Dr. Yu raised the disease phenotyping speed of i2b2 from 1-2 diseases per year to over 1000 per year, and the series of papers have been awarded Editor’s Choice by the Journal of the American Medical Informatics Association, Best Paper by the 2019 Yearbook of the International Medical Informatics Association, and published as a standard biological method in Nature Protocols. After returning to China, Dr. Yu has led a team to develop a series of technologies for Chinese EHR analysis and healthcare artificial intelligence, including high-throughput knowledge graph development, unsupervised medical term discovery, and machine translation for medicine.
