Doctoral Thesis Oral Defense - Meng-Chieh (Jeremy) Lee

— 4:00pm

Location:
In Person and Virtual - ET - Mauldin Auditorium, Newell-Simon 1305 and Zoom

Speaker:
MENG-CHIEH (JEREMY) LEE , Ph.D. Candidate, Computer Science Department, Carnegie Mellon University
https://mengchillee.github.io/

Explainable Mining of Graphs and Time Series: Algorithms and Applications

Given a social network graph, how can we predict connections between users and determine whether they are based on shared hobbies or common friends? Given a database containing molecular graphs, how can we determine whether the graphs inhibit HIV replication based on substructures they frequently share? Similarly, in time series data from EEG recording, how can we identify seizures and explain why they are considered abnormal? Although recent machine learning methods have shown improved performance, many remain black-box models, making explainability challenging. This leads us to explainable artificial intelligence (XAI), which offers valuable insights through its explanations and is more practical for deployment in real-world applications.

In this thesis, we focus on developing explainable machine learning methods tailored for graphs and time series. Each method we propose is either inherently explainable, or designed to automatically provide data analysis or justification for its decisions. In each part, we present effective and general algorithms, and explore a broad range of applictions.

Thesis Committee

Christos Faloutsos (Co-chair)
Leman Akoglu (Co-chair)
Geoffrey Gordon
Nina Mishra (Amazon)

In Person and Zoom Participation.  See announcement.

Add event to Google
Add event to iCal