What I Did
I worked as a research assistant from June 2020 to July 2021
In this role I conducted research on using Transformer models to score the importance of different videos and create summaries of those videos from those scores.
During this research I explored different Transformer model architectures, different input features, different ways of combine those features and more. Through this work I learned a lot about how the attention mechanism in deep learning works, Transformers work, and a variety of image and video processing techniques.
My work here allowed me to expand to other areas where Transformers are used, specifically allowed me to transition over to NLP work.