Title: Virtual human motion generation with machine learning

Abstract: Generating human-like motions has always been a compelling topic of many research fields, including robotics and computer animation. Due to the complex structure and the high-dimensional motion capacity of the human body, approaching the problem with optimization and machine learning techniques dates back to decades ago. And today, with more powerful processors and the emergence of deep learning techniques, better results are achieved. My main aim in this talk is to introduce several ways of employing machine learning techniques for virtual human motion generation. To this end,  I will mention some of our research studies for generating motions like catching, walking with a desired style and avoidance.

Biography: Zümra Kavafoğlu received her M.Sc and Ph.D. degrees from Hacettepe University, Department of Mathematics, with interdisciplinary research studies covering Computer Science topics. During her M.Sc. studies, she performed research on computational geometry and scientific visualization as a part of the TÜBİTAK project “Development of Mine Automation and Design Software”.  Her Ph.D. research focused on physics-based biped motion controllers, as a part of the TÜBİTAK project “Data Driven Virtual Human Animation”. From December 2014 to June 2015, she was a visiting scholar in Utrecht University(the Netherlands), Department of Information and Computing Sciences. Currently, she is an assistant professor in Hacettepe University, Informatics Instittute, Computer Graphics Department. Her current research interests consist of computer graphics and animation, especially physics-based human animation with machine learning and optimization methods, as well as artificial intelligence for games.