For the second year, MSOE students have been challenged by Dr. Dwight Diercks ’90, MSOE Regent, to utilize Rosie the Supercomputer in the Rosie Supercomputer Super Challenge. Five teams have been selected as finalists for this year’s competition competing for a grand prize of $5,000 and an NVIDIA GPU.

The challenge posed to students is to solve an interesting problem, answer a difficult question or go beyond their course work or improve an existing process using Rosie.

Finalists include:

• Light Transport Simulation | Jonathan Paulick and Jonathan Phung
  This group worked to improve the industry-standard method of light transport that is frequently used in biomedical optics through GPU acceleration in order to improve the overall speed of performance.
• Tornado Detection | Jordan Schlick and Benjamin Leisher
  By detecting tornadoes in raw volumetric radar data, this group hopes that this method will help speed the process of issuing warnings in the event of a tornado to help save lives. Utilizing Rosie, this group created a c++ python module using radar data, historical location data and time data to pinpoint a locations likelihood of a tornado.
• CT from X-rays | Ben Paulson, Joshua Goldshteyn, Sydney Balboni, John Cisler, Andrew Crisler, Natalia Bukowski, Julia Kalish and Theodore Colwell
  Building upon existing research, this group looked to explore how orthogonal X-ray images could be converted into a CT image in order to reduce costs and exposure to radiation that a patient experiences during a CT scan.
• Music Transformers | Michael Conner and Jonny Keane
  This group looked at the automated state of the art music transformer model created by Google Brain to improve how effectively the transformer generates its own music composition.
• Prostate Segmentation | Nigel Nelson
  Diagnosing prostate cancer is most often done using biopsy. Automating MRI diagnostics to remove interobserver variability is a far less invasive method of diagnosing patients compared to biopsy. Nelson worked to create an MRI-based cancer semantic segmentation model that would predict the presence of cancer using the less invasive method.

These students will present before a panel of judges which include Diercks, MSOE Regent Nick Haemel ’02, and Dr. Derek Riley, MSOE computer science program director. Presentations will be held in the Diercks Hall NVIDIA Auditorium on Thursday, April 20. An open poster session will begin at 4 p.m. with formal presentations to follow at 4:30 p.m. Members of the MSOE community are invited to attend the event in person. Livestream will also be available for those unable to attend on campus.

To learn more about the Rosie Supercomputer Super Challenge, visit msoe.edu/rosie-challenge

Watch the event via livestream: http://bit.ly/3GA7EMI