The 2021 Asiagraphics (AG) Young Researcher Award was presented to Dr. Yuki Koyama from the National Institute of Advanced Industrial Science and Technology (AIST), Japan. The winner of this award was selected by the award jury chaired by Prof. Ming Lin (UMD College Park) and Prof. Leif Kobbelt (RWTH Aachen).
Dr. Yuki Koyama obtained his Ph.D. degree in 2017 from the University of Tokyo. From April 2017, he has worked as a Researcher at the National Institute of Advanced Industrial Science and Technology (AIST), Japan. His research interests are at the intersection of computer graphics and human-computer interaction (HCI), with an emphasis on applying computational techniques to various design problems.
One of his most important contributions is the development of human-in-the-loop Bayesian optimization (BO) techniques and the application to parametric design problems in computer graphics. BO has been a recent hot topic in the machine learning community as a general method of sample-efficient black-box optimization. His research extends BO in an original way for human-in-the-loop settings to apply it to visual design problems, where designers need to determine which design is subjectively preferred. This work was published at SIGGRAPH 2017, and he later named this powerful approach as preferential Bayesian optimization (PBO) and is actively exploring its unrevealed potential from various viewpoints. For instance, he already proposed a novel variant of PBO at SIGGRAPH 2020, where a tailored user interface is effectively combined with PBO to make the optimization even more efficient.
He has also made considerable contributions to human-in-the-loop optimization other than BO. For instance, his co-authored paper published at SIGGRAPH 2020 proposed differential subspace search, a human-in-the-loop optimization algorithm to efficiently explore the high-dimensional latent space of deep generative models; this paper provides a general approach to the grand challenge of providing means of searching the latent space for the desired expression. Besides, the interactive design systems he developed (e.g., the ones published at SIGGRAPH 2014, UIST 2014, CHI 2016, and CHI 2018) involve interactive optimization in creative ways to enhance designers’ interactive experiences.
He has also worked on design problems on personal fabrication. He worked on computational design of 3D-printable functional objects and published a paper at SIGGRAPH Asia 2015, which has been cited by both computer graphics and HCI papers, indicating its interdisciplinary impact. He also has computational fabrication papers at SIGGRAPH 2014 and CHI 2021.
