Koichi Namekata

I am a final year undergraduate student at the University of Toronto, majoring in Computer Science and Statistics.

My research interests broadly span Computer Vision, Computer Graphics, and Machine Learning with focus on developing generative models and utilizing them for solving various downstream applications. In the past, I actively participated in competitive programming contests, where I won Bronze and Silver medals (1st place among the silver medalists) at the International Olympiad in Informatics. Please see below for more details.

LinkedIn  /  Google Scholar  /  Twitter  /  Email

Publications

The topics of my previous research include Diffusion Models, GANs, Vision-Language Models, Representation Learning, and Unsupervised Learning.

SG-I2V is a framework that enables zero-shot trajectory control in image-to-video generation by relying solely on the knowledge present in a pre-trained image-to-video diffusion model.

EmerDiff is an unsupervised image segmentor solely built on the semantic knowledge extracted from a pre-trained diffusion model. Our experiments suggest the presence of highly accurate pixel-level semantic knowledge in diffusion models.

Projects

High school project funded by Mitou Junior program. Won Minister of State for Science and Technology Policy's Awards at Japan Students Science Awards (日本学生科学賞) in 2019.

Education

Awards

I borrowed the template from Jon Barron.