About the Project
Established methods for robotic control are usually inflexible in adapting to new tasks. Recently, deep neural network based methods for interactive control, termed reinforcement learning, have shown promise in self-learning to solve tasks. However, they require a huge number of, often random, interactions with the environment for each new task. On the contrary, human brains learn models of their bodies and the environment to efficiently predict and plan their decisions and movements, and can adapt these online.
This project seeks to distill and improve diverse advances in cognitively-inspired model-based reinforcement learning to enable robots to self-learn new tasks and adapt to perturbations fast and flexibly. Specifically, the project will research and create a neural-based architecture that is able to learn a multi-level model of the robot and its environment, and employ this model for planning and executing actions. This architecture will enable a robot to self-learn to attain goal states, via planning at a higher, human-interpretable level on its internal model with minimal real-world interactions, and also to adapt online. The student will benchmark the architecture on a series of common tasks, with the aim of deploying these on real-world industrial applications.
The PhD student will be based in the Department of Computer Science at the University of Sheffield, and will be guided by an interdisciplinary collaboration between Dr Aditya Gilra and Dr James Law. Dr Gilra is a Lecturer in the Machine Learning Group, and affiliated to the Neuroscience Institute, researching at the intersection of computational neuroscience and machine learning. Dr Law is a Senior Innovation Fellow in the Complex Systems group and the Advanced Manufacturing Research Centre (AMRC), Director of Innovation and Knowledge Exchange at Sheffield Robotics, and brings expertise in cognitive developmental robotics and industrial human-robot collaboration. The student will have access to world-class facilities in High Performance Computing and Robotics hosted by the University. There are also ample opportunities for self-development in teaching and supervision, and via activities / events of the various affiliated departments as well.
To apply for the project, applicants need to apply directly to the University of Sheffield using the online application system. Complete an application for admission to the standard Computer Science PhD programme
Applicants will require a Master’s upper second class honours (or exceptional Bachelor’s) degree in a quantitative field, with excellent mathematical and programming abilities and a demonstrable specialization in machine learning, robotics, control theory, computational neuroscience, or similar areas.
If English is not your first language, you must have an IELTS score of 6.5 overall, with no less than 6.0 in each component.
Candidates will be interviewed on a rolling basis as the applications come in. Email email@example.com for further details
Why not add a message here
Based on your current searches we recommend the following search filters.
Based on your current search criteria we thought you might be interested in these.