Mevlana Gemici, PhD student, won the Best Paper Award at IROS'14 Cognitive Robotics for his work on learning online unsupervised representation from haptic/tactile data using non-parametric methods. Ashutosh Saxena co-authored the paper.  The new algorithm performs belief space planning on the learned representations, and allowed the robots to perform tasks as complicated as preparing salads by cutting 10s of varieties of food items.  Now Robo Brain will have haptic/tactile information about the world as well!