The Delft Center for Systems and Control and TU Delft Robotics Institute will start the project:
Deep Learning for Robust Robot Control
Project description:
While robots can flawlessly execute a set of commands to achieve a task, these commands are mostly encoded by hand. There is a need for effective learning methods that can deal with the uncertainty in the robot's environment, in particular when only broad goals are specified, and the learning algorithm has to learn motor commands to achieve these goals. This typically involves reinforcement learning (RL). However, current RL for robotics tasks relies on ad hoc function approximators and is typically not robust to changes in the task, environment, or robot uncertainty (compliant robot actuators, or wear and tear). The aim of this project is to integrate two emerging notions in order to make reinforcement learning for robot control more robust and efficient: dynamic feedback control policies for robust control combined with deep neural networks to learn low-dimensional parameterisations of such control policies. This approach promises a generic and robust approach to reinforcement learning for robotic control. The project is a collaboration between Delft University of Technology and the Centre for Mathematics and Computer Science (CWI) in Amsterdam, and is being funded by the Natural Artificial Intelligence programme of the Netherlands Organisation for Scientific Research (NWO). The supervisors and promoters are Profs. Robert Babuska and Karl Tuyls (TU Delft), and Dr. Sander Bohte (CWI) is an advisor on the project.
Description
We are looking for a candidate with an MSc degree in systems and control, applied mathematics, artificial intelligence or machine learning, as well as a strong interest in robotics. Additional experience in the use of deep learning neural networks and/or robotics is an asset. The candidate must have strong analytical skills and must be able to work at the intersection of several research domains (control, machine learning, robotics). A very good command of the English language is required, as well as excellent communication skills.
Nr of positions available : 1
Research Fields
Technology
Career Stage
Early stage researcher or 0-4 yrs (Post graduate)
Experienced researcher or 4-10 yrs (Post-Doc)
More Experienced researcher or >10 yrs (Senior)
Research Profiles
Not defined
Benefits
TU Delft offers employees an attractive benefits package, including a flexible work week and the option of assembling a customised compensation and benefits package (the 'IKA'). Salary and benefits are in accordance with the Collective Labour Agreement for Dutch Universities.
As a PhD candidate you will be enrolled in the TU Delft Graduate School. TU Delft Graduate School provides an inspiring research environment; an excellent team of supervisors, academic staff and a mentor; and a Doctoral Education Programme aimed at developing your transferable, discipline-related and research skills. Please visit www.phd.tudelft.nl for more information.
Submit you application to Prof. Robert Babuska and Prof. Karl Tuyls (email: r.babuska@ tudelft.nl, k.p.tuyls@tudelft.nl) before November 30th 2014. Include a cover letter along with a detailed curriculum vitae, a separate motivation letter stating why the proposed research topic interests you, electronic copies of publications (if applicable), the summary of your MSc thesis, your MSc and BSc course programs and the corresponding grades, names and addresses of two to three reference persons, and other information that might be relevant to your application. The project is expected to start on January 1st 2015.
Benefits
0 - 2664
Requirements
| Technology |
| ENGLISH |
| Good |
| Technology |
| University Graduate |
| 1 |