Exploring Machines that Understand Videos

Photo of Myself

Author: Matt Couts | Major: Computer Science | Semester: Spring 2024

My name is Matt Couts and I am an honors student in the College of Engineering studying Computer Science. With the help of my mentor, Dr. Khoa Luu, I conducted research in machine learning for video understanding.

Throughout the Spring 2024 semester, I researched machine learning under the supervision of Dr. Khoa Luu. The focus of my research was in building machine learning models that are able to converse about the contents of a video via a chatbot. Essentially, my aim is to extend the capabilities of large language models, which are limited to text, to the video modality. This development has significant implications across various fields, including education, medicine, and security.

Excited by the explosive growth of machine learning in the Summer of 2023, I was eager to get involved in the field. After reading about his impressive accomplishments, I reached out to Dr. Luu who agreed to guide my research. I’ve been a proud member of his Computer Vision and Image Understanding lab since then.

In the Computer Vision and Image Understanding lab, I am surrounded by so many intelligent individuals, many of whom are PhD students. Getting to know some of the other students and what they are currently researching has only made me more excited about machine learning. Observing the research of others in the lab has taught me a great deal about my research process and how to improve it. Throughout my research, I worked closely with Pha Nguyen, one of Dr. Luu’s PhD students. From the day I stepped into the lab, Pha has been extremely helpful. He points me to several great learning resources related to my topic and guides my brainstorming sessions with all of the relevant literature.

This semester of research has taught me so much. First of all, I’ve learned how important it is to be methodical in my approach to research. Having a plan and working hard to stick to it keeps your thoughts organized, your experiments consistent, and your confidence high. I found that documenting my steps and findings is crucial when trying to organize them into a paper. Proper documentation is also needed to ensure my experiments can be repeated/criticized by others.

Machine learning is an exciting topic, but it is also very challenging and competitive. There were times this semester where I doubted my capability to conduct this research. The theory behind machine learning is dense and hard for me to grasp fully. I often have to remind myself of the intuitions behind what I am working on, which can be discouraging when talking to others who seem to understand it all. Additionally, the programs I interacted with are complex and often lack thorough documentation. So when things went wrong, as they often did, I felt stranded trying to troubleshoot.

Still, I made it through my first semester of research and am looking forward to the next. Dr. Luu was a great encouragement when I was down on myself. He constantly reassured me that struggling with difficult concepts is a normal part of the research process, which helped me maintain my motivation. Pha ensured that I always had someone to turn to when I needed help understanding. I am grateful for the collaborative and supportive environment in the CVIU lab, which has made a significant impact on my learning experience.

The Honors College Research Grant has provided me with invaluable research experience and has already taught me so much. Through this experience, I have gained a deeper understanding of the research process, from formulating hypotheses and designing experiments to analyzing results and presenting findings. The skills and knowledge I have gained will definitely help me as a student and as a future professional.