Toshiya Hachisuka joined the David R. Cheriton School of Computer Science as an Associate Professor in September 2020. Previously, he was an Associate Professor at the University of Tokyo, where he led his research group on computer graphics. Before coming to the University of Tokyo, he was an Assistant Professor from 2011 to 2014 at Aarhus University in Denmark.
Toshiya received a PhD in Computer Science from the University of California at San Diego in 2011 and a BEng from the University of Tokyo in 2006. His research interests include light transport simulation, computational statistics, numerical analysis, and physically based animation. He has a number of publications spanning these topics in the area of computer graphics.
Tell us a bit about yourself.
I grew up in Japan. After I finished my undergraduate degree at the University of Tokyo, I moved to the United States for graduate studies. I did my PhD at UC San Diego, under the supervision of Henrik Wann Jensen.
After my PhD, when I was looking for work, Henrik — who is originally from Denmark — connected me with a university in Denmark called Aarhus University. I luckily got my first academic job there as an Assistant Professor in the Department of Computer Science. I hadn’t thought about moving to Denmark when I started my PhD studies, but once I moved there I really got to love the city of Aarhus and Aarhus University.
I was thus planning on staying there for many years, but I got an opportunity to move back to Japan as a Junior Associate Professor at the University of Tokyo, and later promoted to an Associate Professor.
I spent about six years at the University of Tokyo, thinking again of staying there for many more years, but life can be unexpected. I decided to come back to North America — to Canada this time — and take a position at the University of Waterloo. Relative to other professors, I guess I have moved long distances for my education and academic positions — from Japan to the United States, then to Denmark to Japan again, and now to Canada — but I’ve always enjoyed the places I’ve lived, studied and worked.
When did you become interested in computer science and computer graphics?
My father was an early adopter of personal computers in Japan. When I was eight years old, we already had a personal computer at home, but back then a personal computer was essentially a device you had to program yourself. I wanted to use the computer, but to do that I had to program it, which was a natural (or forced) introduction to programming. Like many kids in Japan at the time, I also liked and I still do like playing video games. I thus did a lot of things with graphics to learn programming — drawing images, moving characters around — and that’s what got me into computer graphics.
What attracted you to the Cheriton School of Computer Science?
I have been collaborating with several academics in Canada for quite a few years now, most notably Derek Nowrouzezahrai, an Associate Professor at McGill University. Derek actually did his undergraduate degree at the University of Waterloo and I heard many good things about Waterloo from Derek and from others.
Among many other good things, I often heard that Waterloo’s students are impressively bright and strongly motivated. I also heard about how easy it is to set up a collaboration with industry here, which is good for research in computer graphics. All this made me very interested in seeing what’s happening at Waterloo.
My research brings together mathematics, physics, and computer science. It probably does not exactly fit within “traditional” computer science such as algorithms and computer systems, so to speak. Usually, university curricula in computer science are centred on traditional computer science, and students there may not have a strong background or interest in mathematics and physics.
But at Waterloo, computer science is part of the Faculty of Mathematics, and many of our computer science students have strong interests and abilities in math and physics. This rare combination of mathematics and physics with computer science leads to a pool of exceptional students here, the kind who would also be interested in the type of research I love.
I also recently started doing some work in quantum computing and its applications to computer graphics. All these things — Waterloo’s strength in computer science and math, the quality of its students, that students have strong interest and ability in math and physics, along with Waterloo’s strength in quantum computing — made the decision to come here very compelling.
Tell us a bit about your research.
My research currently revolves around light transport simulation and relevant numerical methods. Let’s say you give the computer a shape of a certain object and information about what it’s made of and light sources. Say you want to take a photograph of this virtual scene that exists only in a computer. This is actually how photorealistic images you see in movies and video games are generated. Behind what you’re seeing on screen is very complex simulations of light transport, which is exactly my current research focus.
Such a computation of light transport takes a lot of time. Movies studios these days even have their own cluster of computers to simulate light transport to generate their movies. For example, say an artist added a ceiling light to a room and it emits light only towards the floor. The artist probably expects that the ceiling is not completely dark, but somehow lit by the ceiling light. How does this happen? It’s not because the light is coming directly from the ceiling light itself. The light is bouncing off the floor and walls and other objects in the room and they are collectively indirectly illuminating the ceiling. We thus need to simulate light transport in order to figure out illumination on the ceiling in this case.
My research is essentially about making such simulation of light transport more efficient and expressive enough that, even if an artist gave a complex shape of an object and also a complex configuration of light sources, the algorithm can still solve the light transport in a reasonable amount of time. This involves not only making computation itself more efficient, but also reformulating the problem in a mathematically better way such that you can transform new equations into a more efficient computer algorithm.
Light transport simulation is used to create photorealistic images for movies, but it can be used in many other areas — for example, simulating lighting for interior decoration in a new building. Or it can be used to simulate optics in devices such as the headlights of a car or a lighting fixture. The main user of this computer graphics research at the moment is the movie industry, but the same technology and algorithms to simulate light realistically for a movie can be used to simulate light for any object you can think of.
Do you see opportunities for collaborative research?
I think it is very likely that I will collaborate with the members of the School’s Computer Graphics group, especially with Christopher Batty who does complementary work in computer graphics in a way. We both work on simulation of physics, but just from different perspectives to create photorealistic animations using computers.
I’m also affiliated with Computational Mathematics at the University of Waterloo. I haven’t talked to other members yet, but many faculty have a common interest in applying mathematical knowledge to simulate real-world objects and phenomena. I see a lot of potential to work together on all sorts of topics, not just light transport simulation. Faculty there all basically apply mathematical knowledge to computer simulations.
With my students, I’ve started exploring the possibility of using quantum computing in computer graphics, so collaboration with researchers at the Institute for Quantum Computing is something I would like to start soon. If it goes well, maybe we could realize the first computer-generated animation based on quantum computing! While movie studios may not install quantum computers any time soon, this topic is still really interesting to me as it might achieve something that simply hasn’t been possible to do before.
What do you see as your most significant contribution or work to date?
It’s difficult to name a single project or paper, but if I were to do so based solely on the number of citations so far, the work I did during my PhD on progressive photon mapping might be the most significant. This work enabled accurate computation of new types of lighting effects that are possible to simulate on a computer.
This work is now part of software called RenderMan that’s developed by Disney Pixar and used by many movie studios across the world. One of the features of RenderMan includes the work I did as part of my PhD and afterwards.
Who has inspired you most in your academic life?
Again, it’s difficult to limit an answer to just one person, but if I were to name a single individual, it would be my PhD advisor, Henrik Wann Jensen. He had a huge influence on me.
When I started my PhD, I didn’t really think about becoming an academic afterwards, but the way Henrik conducted research and the way he taught his students inspired me to stay in academia.
I should mention that my collaborators and students inspire me a lot. Working as faculty, in a way, is like doing a PhD for life. I’m always learning new things and discovering new things by working with others, and those people keep inspiring me throughout my academic life.
What do you do in your spare time?
I became interested in computer graphics by playing video games as a kid and I still do enjoy playing video games. Sadly, maybe because I don’t have much spare time now, I have a pile of video games at home that are waiting for me to play [laughs]. I often play video games with my wife and that’s always very fun.
My wife and I also enjoy good food. Once the pandemic has settled down we’re looking forward to exploring culinary experiences in Waterloo and Canada. We can enjoy some delivery or take-out, but I think it’s not just about eating tasty food. We are looking forward to enjoying the entire experience — the ambience, the anticipation, and the meal.