Machine Learning & Optimization
The machine learning and optimization group works on novel algorithms that allow a computer or robot to learn to infer the most probable state of the world, and to take optimal actions, given noisy or incomplete information. We focus on developing efficient distributed optimization algorithms that are well-suited to current computing technologies. Because we focus on fundamental algorithms, our work has a wide variety of applications spanning a number of other Disney research areas, including computer vision, robotics, human-computer interaction, graphics and materials research.
(in alphabetical order)
Facial Motion Retargeting with Input-Output Temporal RBMs
Retargeting facial motion from one actor to another is a complex task. Current methods include: Blendshape Mapping, where target face is expressed as a combination of key shapes with the weight given by source data and Geometric Mapping, where expressions of the target face are modelled as a source face offset with respect to the base target geometry. These methods, however, do not take into account dynamic aspects of the facial motion itself, making it hard, for example, to model speech.