GirlHacks Winner Aims to Help Students with Hearing Loss
Anupriya Gotkhindi was listening to a lecture in her machine learning graduate class when she noticed a sign language translator having difficulty trying to decipher the complex issues being taught and the technical terms used by the instructor.
That problem was the inspiration for her team’s project at the recent GirlHacks Hackathon at New Jersey Institute of Technology (NJIT). The event was the first ever hackathon at NJIT specifically designed to promote and encourage women in STEM.
Gotkhindi’s team topped dozens of teams to win the top prize at the event. For their project, “Sign Language Alphabet Prediction,” they used a machine learning model to convert American Sign Language gesture images into text. Their model was trained using more than 3,000 sign language images compiled from public sources. Such a tool would be helpful to a student with hearing loss if he or she needed to present information to the class using sign language.
While GirlHacks was designed to encourage female students in STEM, it was open to all students. Gotkhindi’s team included fellow graduate students Aniruddha Kulkarni and Jalaj Sharma, both studying computer science, and Nitesh Mistry, who along with Gotkhindi studies data science. All four are currently taking the same machine learning class.
Gotkhindi said her team plans to continue working on their project and hopes to pursue their larger idea of developing an app that can be made available publicly to help others coping with hearing loss.
GirlHacks was organized by NJIT’s Women in Computing Society (WICS) after student board member Yashwee Kothari pitched the idea to the club.
“Computer Science is currently a heavily male-dominated field. I went to my first HackNJIT event and saw it was primarily men and just a few women. I wanted to change that and make CS more inclusive,” said Kothari, a sophomore studying computer science.
Her idea blossomed into a two-day event that attracted more than 250 students and was sponsored by ADP, Barclays, Prudential, Facebook, Expel, .tech, Stickermule, Baslsamiq and Temboo.
For the first-place finish Gotkhindi and each of her team members received a drone and a tour of the ADP headquarters.
“By having a hackathon specifically designed for women, we can help eliminate any hesitation that might stop a student from entering or continuing in this field,” said Tejasree Seelam, a senior computer science major and WICS board member who helped organize the event. “Seeing and working with other women who have the same interests and concerns as you can be extremely comforting and goes a long way to maintain their interest in technology.”
According to U.S. Bureau of Labor Statistics, the number of computer science-related research job opportunities will grow by 19% by 2026, yet women only earn 18% of computer science bachelor’s degrees in the U.S. The diversity numbers are slightly better at NJIT’s Ying Wu College of Computing where 20% of students are women.
After seeing the success of NJIT’s first GirlHacks event, Seelam is hopeful that it will become an annual event at the university.
“Having a dedicated and passionate team of women really helped us build a great foundation that can be continued in the future. I hope to see GirlHacks as an annual event on par with HackNJIT – the annual NJIT hackathon.”