Learning and Adaptive Behavior
Japanese scientists are searching for learning techniques that can scale indefinitely. At the University of Tokyo, researchers are using a learning methodology they call interactive teaching to give robots the ability to drive their own development. The robot uses Bayesian Networks to map sensor evidence to behavior, and then assigns each mapping a confidence rating. In the beginning stages, confidence ratings are low and the robot must frequently ask a human trainer for help deciding between competing actions. With practice, the robot requires less intervention from the human trainer until eventually a task can be completed autonomously. As the task changes, the robot can again ask for help. 7
DB, a humanoid specially designed by SARCOS for the Kawato Dynamic Brain Project.
At the University of Southern California, Maja Mataric is working to provide the SARCOS humanoid, DB, with a set of basis behaviors on which developmental learning can build. 4 Imitative learning is then a process of matching perceived behavior to an assemblage of these a priori primitives. On the same robot, Chris Atkeson, Josh Hale, Mitsuo Kawato, Shinya Kotosaka, Frank Pollick, Marcia Riley, Stefan Schaal, Tomohiro Shibata, Gaurav Tevatia, Ales Ude and Sethu Vijayakumar are also working to emulate complex, full-body movement. 5 For insight into human body movement, they use a unique motion capture system called a SenSuit which, when worn as an exoskeleton, allows researchers to record human movement trajectories for shoulders, elbows, wrists, hips, knees and ankles. This data helps to identify the underlying principles that constrain and optimize body movement. Ultimately, these principles will inform the way motion primitives are developed and used by humanoid designers. Currently, researchers have chosen to represent motion primitives using B-spline wavelets -- spikes in the kinematic graphs that characterize a specific joint movement. By providing an efficient way to specify and optimize multi-resolution motion trajectories, B-spline wavelets enable smooth, efficient movement.
Work with humanoid learning also progresses at Michigan State University, where a technique called communicative learning is used to iteratively hone behavior as the humanoid responds to verbal feedback from a human trainer. The foundational principle of this effort is that all human-derived forms of representation bias the system and inhibit the ability of learning to scale. Instead, they wish the humanoid to build layers of control using as little built-in representation as possible. Rather than storing semantic information, the humanoid treats all stimulation as low-level vectors. Thus, the principles that will allow the robot to process and learn from visual stimulation will apply equally well to other capabilities such as object manipulation.