TNJSF JEI Award Winners: Ryan Rong and Curtis Zhou

Ryan Rong and Curtis Zhou are rising seniors at the Peddie School. Their project was focused on creating a machine learning model that could bilaterally translate American Sign Language and English. They were able to do this by proposing a new architecture that used both key point extraction and attention architecture to allow for recognition of sign language. Curtis initially came up with the idea for a sign language translator after seeing one of his friends who is deaf and hard of hearing get into a situation they didn’t want over miscommunication. Going to Ryan, who shared similar concerns, they were able to combine their knowledge in computer science to build their award winning machine learning model.

One of the biggest obstacles they faced outside of building their model was getting everything setup for a public facing website for their project so it could be used by others. During this process Ryan and Curtis learned how to export and host their model, how to send videos to their model to analysis, and how to control the traffic coming to their website. They worked very hard to minimize any delays present so their project could be easily used by others in everyday life. Following the trend of previously featured winners, both Ryan and Curtis noted the importance of being able to communicate their work to a broad audience. This is especially important on machine learning models where ideas are often complex and abstract until compared to something relatable for the audience.

Congrats to Ryan and Curtis on their success and their work on this project. We are excited to see the impact it will have on others!

The material on this page was prepared by Kari Mattison, JEI co-Editor in Chief. Ryan Rong and Curtis Zhou provided the photo and answers to standard questions which was edited lightly for clarity.