Postdoctoral Research Assistant in Decision-making under uncertainty for robots
University of Oxford, Oxford
Postdoctoral Research Assistant in Decision-making under uncertainty for robots
£39424-£47779
University of Oxford, Oxford
- Full time
- Permanent
- Onsite working
Posted 1 week ago, 6 Sep | Get your application in now before you're too late!
Closing date: Closing date not specified
Job ref: 90ef411703d54c0b9407666ada0c1673
Location ref: Oxford
Full Job Description
Grade 7 : £39,424 - £47,779 per annum
We are looking for a post-doctoral research assistant to join the GOALS group (http://ori.ox.ac.uk/labs/goals/) within the Oxford Robotics Institute (ORI), as part of the EPSRC Programme Grant in Embodied Intelligence (https://embodiedintelligence.web.ox.ac.uk) and other follow-on projects.
The successful applicant will be committed to our research on decision-making for autonomous systems. Our work focuses on mission planning under uncertainty, and sequential decision-making more generally. We tackle problems in single and multi-agent systems operating in uncertain environments and around humans. We typically work with Markov decision processes (MDPs), and are interested in both fundamental algorithms for producing policies for rich goal structures in MDPs (e.g. risk, temporal logic, or probabilistic objectives), and modelling robot decision problems using MDPs (e.g. human-robot interaction, safe exploration, long-term monitoring). Our recent research on algorithms has included work on epistemic uncertainty, uncertain MDPs, proactive task allocation, multi-robot coordination, offline reinforcement learning (RL), Bayes adaptive RL, and planning with novel variants of Monte Carlo tree search.
You should possess a relevant PhD/DPhil or be near completion together with relevant experience, including a track record of contributions in decision-making under uncertainty robotic exploration and mission planning under epistemic uncertainty. You will have knowledge or experience in modelling and solving problems under uncertainty, from either an AI, robotics, or verification perspective, and an ability to manage own academic research and associated activities.