CVPR Daily - Wednesday

Jan Kautz is Senior Director of Visual Computing and Machine Learning Research, which effectively means leading the research team at NVIDIA called Learning and Perception. Jan, I understand that you are opening a new lab in Toronto, under the leadership of Sanja Fidler . Yes, we have been engaged with the community in Toronto for a while now and have an office there. We wanted to expand our deep learning and machine learning research and Sanja is a very stellar individual who we are lucky to work with, so we decided that it would make sense for us to open a research lab in Toronto given all the talent there. As far as I know, NVIDIA is a hardware company. Why do you need to hire the best software talent that is around? We shouldn’t call NVIDIA a hardware company any more. We make the hardware, but also software building blocks that are used by our customers and partners. We do both hardware and software, which makes us a platform company. Software has become a very important part of NVIDIA’s business. That is why we do a lot of research on the software side. We still do research on the hardware side as well, but the software side is becoming more important. That is why you are taking the best talent from us! [ both laugh ] How many employees you have in total today? We have roughly 12,000 employees in total worldwide. Can you tell me about your own work? I am part of NVIDIA Research. Its goal is to push the state-of-the-art forward, so that the research has an impact on the community, but of course, ultimately also on NVIDIA. We try to develop new technologies through research that will help NVIDIA directly or one of our partners at some point in the future. NVIDIA works with a lot of partners and often our research is helpful to them. My group specifically, the Learning and Perception Research Group, do a lot of work in computer vision and machine learning. On the computer vision side, we address a lot of the computer vision stack. Going from low-level computer vision, such as image processing and computational photography, all the way up to high-level computer vision such as scene understanding and video understanding. We try to touch on all of computer vision and address areas that we believe can have an impact on us. Jan Kautz - NVIDIA 4 Guest Wednesday