Interests:

Advancing machine reading for precision health by unlocking structured information from text.

Background:

Dr. Poon received his B.S. with Distinction in Computer Science from Sun Yat-Sen University in Guangzhou, China, and his PhD in Computer Science and Engineering from the University of Washington in Seattle, specializing in machine learning and natural language processing (NLP). He joined Microsoft Research in 2011 and is currently the Director of Precision Health NLP.

Research:

Dr. Poon leads Project Hanover at Microsoft Research, with the overarching goal of advancing machine reading for precision health, by unlocking structured information from text. Three representative scenarios stand out.

• Molecular tumor board: A major bottleneck in precision oncology lies in interpreting tumor mutation, which requires curating knowledge from biomedical literature that adds 4000 papers a day.
• Real-world evidence: Drug development has become unsustainable, due to skyrocketing cost of clinical trials. Electronic medical records (EMRs) contain valuable outcome information for post-market surveillance, drug repurposing, and synthetic control. However, manual curation takes hours per patient.
• Clinical trial matching: 20% of trials fail due to insufficient patient recruitment, which relies on word-of-mouth efforts that are hard to scale.

Assisted curation powered by machine reading can drastically accelerate curation efficiency. However, traditional machine reading requires painstakingly annotating many labeled examples, which limits its applicability. Project Hanover overcomes the annotation bottleneck by combining deep learning with probabilistic logic, and by exploiting indirect supervision from readily available resources such as ontologies and existing databases.

Dr. Poon has given tutorials on these topics at top AI conferences such as the Association for Computational Linguistics (ACL) and the Association for the Advancement of Artificial Intelligence (AAAI). His research spans a wide range of problems in machine learning and natural language processing (NLP), and his prior work has been recognized with Best Paper Awards from premier venues such as the North American Chapter of the Association for Computational Linguistics (NAACL), Empirical Methods in Natural Language Processing (EMNLP), and Uncertainty in AI (UAI).

Representative publications:

Google Scholar https://scholar.google.com/citations?user=yqqmVbkAAAAJ&hl=en