Science Success Story
Student Wins Poster Award at the ERN Conference in STEM
Jetstream, Wrangler supercomputers enable winning scientific research
By Harmony Jankowski (IU) and Hannah Rae Remmert (NCSA)
Tenecious Underwood, a senior at Livingstone College in North Carolina, has already had quite a year.
|Tenecious Underwood presenting at SC19.|
In early 2019, Underwood, who always had an affinity for math and computers, was encouraged by a fraternity brother he met at the National Science Foundation's (NSF) Emerging Researchers National Conference in STEM to apply for the Jetstream REU program. This program is hosted at Indiana University (IU) and is made possible, in part, through XSEDE Education Allocations.
Each summer, Jetstream — the NSF's first production cloud computing system which provides on-demand HPC and data analysis resources for research and education co-located at IU and the Texas Advanced Computing Center (TACC) — sets students free in the cloud on projects that capitalize on IU's leadership in fields like bioinformatics, data visualization, and advanced media. Underwood, Evan Suggs (University of Tennessee, Chattanooga), and Eliza Foran (Indiana University, Bloomington) created a workflow to record and identify animal vocalizations, easing the burden of data collection on field biologists.
|Tenecious Underwood and team won first place in the computer science poster category at this year's Emerging Researchers National Conference in STEM for work made possible in part through XSEDE.|
Specifically, they wanted to identify amphibians local to Indiana. They gathered sample calls for four species from the Macaulay Library archive of wildlife sounds at Cornell, and used 85 percent of the data to train three neural networks that could process data collected in the field. In essence, Underwood and his colleagues used machine learning and HPC to solve the problem of identifying frog migration patterns.
The team outlined a proof-of-concept workflow that makes the entire process from gathering to interpreting data, more attainable for researchers.They simulated the data collection process by collecting animal (frog) calls using recording devices and Raspberry Pi's (low-cost, fully functional computers the size of credit cards). The team then fed this data into a database and virtual machine hosted on TACC-allocated XSEDE resources (i.e. Jetstream and Wrangler). Processing the more than 5,000 audio files that were recorded would have taken up to 20 hours on a laptop computer, but using Jetstream, it took minutes.
Watch Tenecious Underwood explain how he and his team used machine learning to identify frog calls here (video via Science Node):
Once the research was done, Underwood's journey took off.
He, Suggs, and Foran presented their work at PEARC19 in Chicago. Then, in November, Underwood presented their work at SC19 in Denver to great acclaim. Most recently, he presented Automatic Recognition of Frog Calls earlier this month at this year's ERN conference earlier, and won first place in the Computational Computer Sciences and Information Management Division of undergraduate student posters category.
Underwood is now preparing to graduate from Livingstone College in just a few months. What's next? Graduate study in computer science at Kentucky State University with a specialization in cybersecurity. Underwood's bright outlook is a credit to his accomplishments.
Original reporting via Indiana University.
At a Glance
- Undergraduate student Tenecious Underwood wins poster award at ERN Conference in STEM using XSEDE Education Allocations.
- Jetstream, Wrangler supercomputers enable winning scientific research.
- Student team created a workflow to record and identify animal vocalizations, easing burden of data collection on field biologists.
- Processing the more than 5,000 audio files that were recorded would have taken up to 20 hours on a laptop; but using Jetstream in took minutes.
- Underwood is now preparing for graduate study in Computer Science at Kentucky State University.