Professor Luke K. McDowell

lmcdowel@usna.edu
(note the unfortunate lack of a final 'l' in that email address)

Mail:
    Chair, Dept. of Computer Science
    United States Naval Academy
    597 McNair Road
    Annapolis, MD 21402

Phone:
    410-293-6801
Fax:
    410-293-2686
Office:
    Hopper 439

 


Teaching: Current Courses

Research

My research focuses on how to mine useful information from networks of linked data. There are many examples of such data networks including:

Much of my work has focused on different techniques for "collective classification", where we wish to make predictions about the people/pages/nodes in a network, based on the information known about those nodes (e.g., their characteristics) as well as the information known about the other nodes that they link to. Often, using such information can lead to much more accurate predictions. This is a specific problem in the field of statistic relational learning.

I have been at USNA since 2005, was promoted to Associate Professor in 2010, and to Professor in 2016. During 2011-12, I had a great year visiting the Distributed Information Systems Laboratory of Prof. Karl Aberer in Lausanne, Switzerland. It's not often up to date, but you can check out my CV.

My Ph.D. in Computer Science is from the University of Washington. My advisors were Oren Etzioni and Alon Halevy. Earlier, I studied Electrical Engineering at Princeton University, advised by Margaret Martonosi.

Data for Link-based Classification problems

The data for several of our papers is available (along with some code).

Collective Inference (learning in graph-based/networked data) (slides from a recent overview talk)

Predictions from Mobile Data and Twitter

Information Extraction from the Web (for the Semantic Web)

Dimensionality Reduction

Mangrove and Semantic Email: Making the Semantic Web Real

Computer Architecture

Computer Science Education

Teaching: Past Courses

Diversions