Jeremy Hyrkas

Contact

E-mail: My-last-name AT cs DOT washington DOT edu
Office: CSE524

Basic

I am a fourth-year Ph.D. student in the Department of Computer Science & Engineering. I am interested in researching database systems, machine learning methods, and new methods for getting the two to play along. I am also specifically interested in helping the physical sciences leverage these techniques. I am currently pursuing these interests with my advisor, Bill Howe; I am also co-advised by Ed Lazowska.

I earned my Bachelor's of Science in Computer Science with a Minor in Mathematics at Colorado State University in Fort Collins, CO, and graduated Summa Cum Laude. I studied computational approaches to understanding alternative-splicing in plants with Asa Ben-Hur

Research

I am currently working with a group of oceanographers at UW on analyzing data from their SeaFlow cytometer project. A single cruise with the cytometer allows scientists to generate gigabytes of data, and as more and more data is accumulated, the techniques used to identify phytoplankton groups and associated trends will have to be reconsidered on account of scale. We believe that a database system is the right way to store and analyze this scale of data, and we hope to incorporate statistical modeling and machine learning methods directly into such a system.

This work is joint work with Daniel Halperin and the Myria project in the database group at UW CSE.

Publications

Scalable clustering algorithms for continuous environmental flow cytometry.
Jeremy Hyrkas, Sophie Clayton, Francois Ribalet, Daniel Halperin, E. Virginia Armbrust, and Bill Howe.
Bioinformatics.
link

Gaussian Mixture Models Use-Case: In-Memory Analysis with Myria.
Ryan Maas, Jeremy Hyrkas, Olivia Grace Telford, Magdalena Balazinska, Andrew Connolly, Bill Howe.
3rd IMDM Workshop at VLDB 2015. link

Time-Varying Clusters in Large-Scale Flow Cytometry.
Jeremy Hyrkas, Daniel Halperin, and Bill Howe.
27th Conference on Innovative Applications of Artificial Intelligence 2015. pdf

Compiled Plans for In-Memory Path-Counting Queries.
Brandon Myers, Jeremy Hyrkas, Daniel Halperin, and Bill Howe.
1st IMDM Workshop at VLDB 2013. pdf

Personal

Music is one of my biggest passions; I can frequently be found in a record shop or at a concert, and I've been known to play in a band from time to time. I also like to consider myself somewhat of a film and television buff.