Dae Hyun Lee

Education

  • Doctor of Philosophy (Sept. 2014 - Dec. 2019)
    • University of Washington, School of Medicine
    • Department of Biomedical Informatics and Medical Education
    • Research Advisor: Dr. Meliha Yetisgen and Dr. Eric Horvitz
    • Dissertation: Predictive Approaches for Acute Adverse Events in Electronic Health Records
  • Bachelor of Engineering (Mar. 2006 - Aug. 2014)
    • Yonsei University, College of Engineering
    • Department of Computer Science
    • Mandatory Military Service: Nov. 2009 - Sep. 2011

Research interest

  • Automatic Speech Recognition
  • Building prediction model on clinical data using machine learning
  • Extracting novel concepts indicating patient deterioration from clinical notes using natural language processing
  • Applying data-driven approach to discover treatable complication during hospital stay, failure-to-rescue

Research Experience

  • Dec. 2019 - Present: Applied Scientist, Amazon Transcribe

    Responsible for improving Automatic Speech Recognition stack

  • June, 2018 - Sept. 2018 : Applied Scientist Intern, Amazon

    Focused on improving cost-effectiveness on intent annotation when introducing new features into existing Alexa system

  • June. 2017 - Sept. 2017 : Research Intern, Microsoft Research (Mentor: Eric Horvitz)

    Focused on predicting the onset of acute organ failure using different annotation strategies.

  • June. 2016 - Sept. 2016 : Research Intern, Microsoft Research (Mentor: Eric Horvitz)

    Continued to work on predicting ICU mortality using PhysioNet ICU Challenge data. Built the model to predict ICU in-hospital mortality. The model infers parameter which governs decaying risk of survival during ICU stay and presented the contribution of clinical variables for the prediction.

  • Sept. 2015 - June. 2016 : Predoctoral Research Assistant, Mooney Group, University of Washington

    Focused on improving current newborn screening method with predicted pathogenicity from their genetic sequence. Built a prediction pipeline by integrating genetic knowledge bases.

  • July. 2015 - Sept. 2015 : Research Intern, Microsoft Research (Mentor: Eric Horvitz)

    Exposed to applying machine learning on clinical data. Worked on understanding clinical data and discovered the atypia of clinical variable could be useful feature predicting the risk of clinically adverse events.

  • Sept. 2014 - July 2015 : Predoctoral Research Assistant, Sauro Lab, University of Washington

    Developed a plugin for Tinkercell to adopt new SBOL(Synthetic Biology Open Language) standard(version 2.0)

  • Sept. 2013 - Sept. 2014 : Researcher, Biomedical Informatics Lab, Seoul National University College of Medicine, Seoul, Korea

    Developed agent apps in Health Avatar (a software platform for personalized management of health information) as a main developer. Also, Worked for integrating various pathway database (WikiPathway, KEGG).

  • June 2012 - May 2013 : Undergraduate Intern, Embedded and Bio database Laboratory, Yonsei University, Seoul, Korea

    Experienced the field of bioinformatics for the first time. Worked on developing the algorithm for partitioning Protein-Protein Interaction (PPI) network.

  • May 2008 - August 2008 : Distributed System Researcher,Helsinki Metropolia University of Applied Science, Espoo, Finland, FKII, Finland-Korea Internship Exchange Program.

    Took in charge of researching preliminary research on computational cluster. Benchmarked the performance between single machine and distributed system using Message Parsing Interface(MPI)

Publications

  • Lee DH, Yetisgen M, Vanderwende L, Horvitz E. Predicting severe clinical events by learning about life-saving actions and outcomes using distant supervision. Journal of biomedical informatics. 2020 Jul 1;107:103425.
  • Lee D and Horvitz E. Predicting mortality of intensive care patients via learning about hazard. Proc 31th Conf Artif Intell (AAAI 2017). 2017.p. 4953-4954
  • Adhikari AN, Wang Y, Gallagher R, Zou Y, Sunderam Y, Shieh J, Chellappan A, Bassaganyas L, Cai B, Chen F, Freedman G, Koenig BA, Kvale M, D. Lee, Vaka D, Zerbe B, Mooney SD, Srinivasan R, Kwok PY, Puck JM, Brenner SE, The NBSeq Project. Exome sequencing of infant dried blood spots identifies three-quarters of metabolic disorders found by newborn screening, indicating limits to exomes in both newborn screening and diagnostic testing. ASHG 2016 Annual Meeting. 2016
  • Burke JL, Lee D, and James RC. The EH Tracker: Using Dynamic Environmental Health Data for Improved Personal Health Decision-Making. AMIA Annu Symp Proc. 2015
  • Ahn J, Lee DH, Yoon Y, Yeu Y, Park S. Improved method for protein complex detection using bottleneck proteins. BMC Med Inform Decis Mak. 2013;13 Suppl 1(Suppl 1):S5. doi:10.1186/1472-6947-13-S1-S5.
  • Ahn J, Lee DH, Yoon Y, Yeu Y, Park S. Protein complex prediction via bottleneck-based graph partitioning. Cikm. 2012:49. doi:10.1145/2390068.2390079.

Poster & Presentation

  • Lee D, Yetisgen M, Horvitz E. Learning about Life-Saving Interventions to Predict the Risk of Acute Organ Failures. 2018 AMIA Annual Symposium Podium Abstract. San Francisco CA, USA
  • Lee D, Yetisgen M, and Horvitz E. Predicting Next Day Need for Life-Saving Interventions in Intensive Care Patients. 2018 SCCM Annual Congress. San Antonio, TX.
  • Lee D and Yetisgen M. Clustering Vital Sign Observations Using Unsupervised Random Forest. 2017 AMIA Annual Symposium Podium Abstract. Washington DC. USA
  • Lee D and Horvitz E. Predicting mortality of intensive care patients via learning about hazard. Proc 31th Conf Artif Intell (AAAI 2017). Poster Presentation & Oral Presentation
  • Lee D, Identifying Failure to Rescue using machine learning approach, 2017 Amazon Graduate Research Symposium, Seattle, WA
  • Lee D, Cai B, Mooney S. General Prediction Pipeline for Mendelian Disorders using Whole-Exome Sequencing Data. Human Single Nucleotide Polymorphisms & Disease, 2016 Gordon Research Conference. South Hadley, MA.
  • Burke JL, Lee D, and James RC. The EH Tracker: Using Dynamic Environmental Health Data for Improved Personal Health Decision-Making. 2015 AMIA Annual Symposium Student Design Challenge. San Francisco, CA.
  • Lee D, James R. Using Pathway as an anchor for data collection. 2015 CCD Summer Short Course. Pittsburgh, PA.

Teaching

  • Teaching Assistant, CSE 546, Machine Learning, Autumn 2016
    • Instructor : Sham Kakade
  • Teaching Assistant, CSE 446, Machine Learning, Winter 2017
    • Instructor: Emily Fox
  • Teaching Assistant, CSE 446, Machine Learning, Spring 2017
    • Instructor: Ali Farhadi
  • Teaching Assistant, CSE 546, Machine Learning, Autumn 2017
    • Instructor: Kevin Jamieson
  • Teaching Assistant, CSE 599, Adaptive Machine Learning, Winter 2018
    • Instructor: Kevin Jamieson
  • Teaching Assistant, CSE 573, Artificial Intelligence I, Winter 2019
    • Instructor: Hannaneh Hajishirzi
  • Teaching Assistant, CSE 599 D1, Advanced Natural Language Processing, Spring 2019
    • Instructor: Hannaneh Hajishirzi

Academic Awards

  • Best Presenter Award, Failure to Rescue: Predicting Patient Deterioration using Machine Learning, Global Top Talent Forum by Hyundai Motor Company, August 2017, San Diego, CA
  • Top 10 Student Abstract Nomination, Lee D and Horvitz E. Predicting mortality of intensive care patients via learning about hazard. Proc 31th Conf Artif Intell (AAAI 2017)
  • Third Prize, The EH Tracker: Using Dynamic Environmental Health Data for Improved Decision-Making of Health, Student Design Challenge, American Medical Informatics Association, November 2015
  • College Honors, Yonsei University, Korea, Spring 2012, Fall 2012
  • Second Prize, ARII - Blog Search Engine Using Coord(NHNDev) and Lucene(Apache), Winter of Code, NCSoft, Korea, March 2009
  • Bronze Medal, Korea Olympiad in Informatics, Korea, 2001
  • First Prize, Korea Olympiad in Informatics, Incheon, 2001, 2003
  • Gold Medal, Korea Olympiad in Informatics, Incheon, 2000, 2002, 2004

Scholorship

  • Hyundai Motor Company Fellowship, January 2018 - December 2019
    • Offer-based fellowship from Hyundai Motor Company. Supports $2,000 per month as stipend.
  • GKS Global Korea Scholarship, September 2014 - August 2016
    • Scholarship from Korean Government. 66 students who are expected to pursue graduate studies in foreign countries were selected for the scholarship. Supported $31,000 for 2014 and $40,000 for 2015
  • JinRee Scholorship for academic excellence, Yonsei University, March 2007, March 2012

Service

  • Volunteer, University of Washington Engineering Discovery Days, 2015
  • Military Service, discharged as sergeant, Korean Augmentation to US Army, 19th Expeditionary Sustainment Command, 8th US Army, Nov. 2009 - Sept. 2011 - ARCOM(Army Commendation Medal)
    • MOVSM(Military Outstanding Volunteer Service Medal)

Language Skill

  • Korean: Native
  • English: Fluent (speaking, reading, writing)
  • Japanese: Intermediate (speaking); Basic (reading, writing)

Other

  • Graduate student member - American Medical Informatics Association - Society of Critical Care Medicine