Photo By Public Affairs Office | Ten Naval Undersea Warfare Center Division Newport engineers and scientists competed...... read more read more
A group of 10 Naval Undersea Warfare Center (NUWC) Division Newport engineers and scientists made their way to the former Brooklyn Navy Yard in New York to compete in Hack the Machine, a Navy-sponsored event attracting students, industry professionals and subject matter experts to help solve the Navy’s foremost digital challenges. More than 600 people from the research and development enterprise competed in the Sept. 6-8 event for cash prizes in three different tracks: data science, cyber hacking and digital manufacturing challenges.
By the end of the event, Division Newport employees made significant contributions on winning teams from the cyber and data science tracks.
The Data Science Team
Four Division Newport employees — James Luther, Ranges, Engineering and Analysis Department; Colby Wilkinson, Undersea Warfare (USW) Combat Systems Department; and Ajaz Baig and David Rubenstein, both from the USW Weapons, Vehicles and Defensive Systems Department — along with three international students from New York University (NYU), a student from the University of Southern Alabama, a neuroscience researcher from NYU and a software engineer from the craft website Etsy, formed “The Jumbos” and worked on the challenge Data Science: Cleared for Takeoff.
Their challenge involved two data sets: a compilation of error codes of F-18 aircraft over the past seven years and a list of maintenance actions during that same time. This was the first time the Navy released this information to the public. Teams had to identify which aircraft had wiring issues. With more than a million error codes in total and about 900 unique codes, combined with hundreds of thousands of maintenance actions, the team had a daunting task. Their plan was to make sense of the data and delegate tasks.
On Day One of the event, Rubenstein created a private page for the team to share information. Meanwhile, Luther downloaded a free trial of a software tool to help the team better visualize the data. He taught himself how to use the program in less than a day.
“[The software] … turned out to be a great data visualization tool,” Luther said.
The team also used other software, as well as visual analysis. By the end of the two days of hacking, The Jumbos took home the top prize, which was $20,000 for the team.
The team also found data trends that the organizers had not discovered. For example, there was a spike in wiring problems in a particular timeframe. Plans are underway to follow up with The Jumbos on some of their findings. Moreover, the team was able to train a machine learning algorithm to classify the type of maintenance action needed based on the frequency of error codes in the previous 30 days.
The point of this challenge was to find predictive data. The team was inspired by working with data on that scale and hopes to bring more of this type of data set analysis to Division Newport.
“Having a data team at Division Newport would be useful to visualize data and look for patterns and correlations,” Luther said.
“Any time there’s an experiment, a simulation, or exercise, there is data collected. People here have the skills to provide insights with state-of-the-art algorithms,” Rubenstein said.
Phil Marcelino, of the USW Weapons, Vehicles and Defensive Systems Department, also competed in the Data Science track on a mixed team of government, industry and academic performers called Sharp Cheddar. That team did not win, but Marcelino learned a lot.
“I walked in not knowing what to expect. I wanted to learn as much as I could and it was a great learning experience,” said Marcelino.
The Cybersecurity Team
The Jumbos weren’t the only NUWC winners at Hack the Machine. Team
| Date Taken: | 10.01.2019 |
| Date Posted: | 10.01.2019 09:44 |
| Story ID: | 344828 |
| Location: | NEWPORT, RI, US |
| Web Views: | 58 |
| Downloads: | 0 |
This work, NUWC Division Newport employees on winning teams at Hack the Machine, by Public Affairs Office, identified by DVIDS, must comply with the restrictions shown on https://www.dvidshub.net/about/copyright.
No keywords found.
LEAVE A COMMENT