[Eecs_mscs] EECS Research Seminar Talk with CS Interview Candidate Dr. Ziyang Song-Deep and Probabilistic Representation Learning for Electronic Health Records

[Hunter, Tiffany]

Title: Deep and Probabilistic Representation Learning for Electronic Health Records
Abstract:
Artificial Intelligence (AI) has achieved remarkable success across a broad range of fields; however, it often remains insufficiently prepared for deployment in high-stakes scenarios like healthcare. Recent advancements in AI for language and vision have increasingly prioritized accuracy over explainability, thereby constraining their practical utility and potentially posing risks to human lives. In contrast, while statistical machine learning models excel in providing interpretable analyses, they often fall short in predictive performance. We aim to explore AI solutions for healthcare applications that combine strong predictive performance with strong explainability. In this talk, I will discuss three key research topics: (1) generative AI for healthcare time series, enabling interpretable trajectory analysis; (2) probabilistic models for medical language modeling, integrating neural networks into inference for greater flexibility; (3) a long-term vision for combining deep learning and statistical machine learning to advance scientific discovery.
Bio:
Ziyang Song is a Ph.D. candidate in the School of Computer Science at McGill University. His research revolves around Artificial Intelligence for health, with a particular focus on generative artificial intelligence and statistical machine learning. Specifically, his research designs AI models with explainability by bridging deep learning and statistical models for healthcare applications. He has published in top-tier machine learning (e.g., ICLR, KDD) and interdisciplinary conferences (e.g., MLHC, ACM-BCB). He also received notable honors, including KDD Health Day Best Paper Award (2022) and ACM BCB Rising Star Award (2024). He also secured Quebec's prestigious FRQNT Doctoral Research Scholarship (2022-2026) for deploying real-world AI systems for at-risk population. He has collaborated with Université de Montréal (UdeM) and conducted research at Centre hospitalier universitaire Sainte-Justine (CHU-SJ), where he designed AI models to identify high-risk patients with interpretable analysis in clinical settings.
________________________________________________________________________________
Microsoft Teams Need help?<https://aka.ms/JoinTeamsMeeting?omkt=en-US>
Join the meeting now<https://teams.microsoft.com/l/meetup-join/19%3ameeting_ZWIyMWJlMGQtMzY2NS00YWEyLThiZmItOTA2MzBmNDg2OTBi%40thread.v2/0?context=%7b%22Tid%22%3a%22f3308007-477c-4a70-8889-34611817c55a%22%2c%22Oid%22%3a%22685c3f4f-29d5-4141-ada5-0fdeab8480e4%22%7d>
Meeting ID: 221 159 600 495
Passcode: iJ2ya2it
________________________________
Dial in by phone
+1 614-706-6572,,426835977#<tel:+16147066572,,426835977> United States, Columbus
Find a local number<https://dialin.teams.microsoft.com/8f5f7319-0053-4423-a154-4f8b6e7fb7dd?id=426835977>
Phone conference ID: 426 835 977#
For organizers: Meeting options<https://teams.microsoft.com/meetingOptions/?organizerId=685c3f4f-29d5-4141-ada5-0fdeab8480e4&tenantId=f3308007-477c-4a70-8889-34611817c55a&threadId=19_meeting_ZWIyMWJlMGQtMzY2NS00YWEyLThiZmItOTA2MzBmNDg2OTBi@thread.v2&messageId=0&language=en-US> | Reset dial-in PIN<https://dialin.teams.microsoft.com/usp/pstnconferencing>
[https://www.ohio.edu/sites/default/files/2018-11/invite_logo_teams.jpg]
If you encounter issues with this meeting, please visit the Help link. If you are not able to resolve the problems, please contact the meeting organizer to let them know you are having difficulty.
Org help<https://www.ohio.edu/oit/services/collaboration/teams/help>
________________________________________________________________________________

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.ohio.edu/pipermail/eecs_mscs/attachments/20250217/4dd2139d/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: text/calendar
Size: 10817 bytes
Desc: not available
URL: <http://listserv.ohio.edu/pipermail/eecs_mscs/attachments/20250217/4dd2139d/attachment.ics>