Currently, I am involved in three European projects at IDSIA:
NanoBioTouch - Europe's most advanced tactile sense research programme.
In this project we apply innovative machine learning methods (e.g. recurrent neural networks) for the investigation of tactile data produced by project partners, such as output from modeled tactile sensory organs, microneurographic readings and micro- and nanosensors. Furthermore, we use a robotic finger and a MEMS touch sensor for active learning to speed up the learning process, and investigate the application of artificial curiosity in robotic platforms. The learned behavior of the robotic finger will be compared with human exploratory behavior in tactile curiosity.
IM-CLeVeR - Intrinsically Motivated Cumulative Learning for Versatile Robots.
In IM-CLeVeR we apply Jürgen Schmidhuber's theory of intrinsic motivation and artificial curiosity in the iCub humanoid robot. My current contribution to this project is to investigate and demonstrate the application of different machine learning algorithms within the framework of intrinsically motivated and cumulative learning.
WAY - Wearable interfaces for hand function recovery. We develop algorithms that facilitate control of hand prosthesis and hand exoskeletons through brain-computer interfaces. The project focuses on the practical application of hand assistive devices, and hence, has a strong involvement of patients in clinical trials.
