Don't miss our weekly PhD newsletter | Sign up now Don't miss our weekly PhD newsletter | Sign up now

  Machine-Learning-Based Control of Complex Networks - exploring interactive machine-learning architectures to enable transparent knowledge exchange between artificial intelligence and human experts


   Department of Electrical and Electronic Engineering

This project is no longer listed on FindAPhD.com and may not be available.

Click here to search FindAPhD.com for PhD studentship opportunities
  Dr Y Gu  Applications accepted all year round

About the Project

Applications are invited for a PhD studentship in Machine-Learning-Based Control of Complex Networks. The work will be based within the Control and Power Group in the Department of Electrical and Electronic Engineering. The student will be supervised by Dr Yunjie Gu and co-supervised by Dr Tae-Kyun Kim. The studentship will start as soon as possible from September 2018.

Electrical power networks are increasing in complexity and uncertainty due to the growth of distributed and intermittent renewable energy, calling for new automation tools for real-time network control. Machine-learning-based control is a promising solution. Simulation-based reinforcement learning proves to be capable of generating control policies in complex and uncertain environments with superhuman performance, but still suffers from very low interpretability, which hinders human-machine interaction and poses a significant barrier to industrial application. This studentship will explore interactive machine-learning architectures to enable transparent knowledge exchange between artificial intelligence and human experts. Possible directions include but are not limited to case-based reasoning, Bayesian inference, visualisation, localisation, and semantic analysis. The work will focus on general mechanisms in machine learning and takes the application in power networks as a by-product.

Applications are invited from candidates with (or who expect to gain) a Distinction/First-Class Master’s degree or an equivalent degree in Engineering, Mathematics, Physics or a related discipline. A strong background in either machine learning or control engineering is required, and skills in Python and knowledge of power networks would be beneficial.

For a description of the Control and Power Group please visit our website at
http://www3.imperial.ac.uk/controlandpower.

Informal enquiries and requests for additional information for this post can be made to Dr Yunjie Gu by email at [Email Address Removed].

Applications will be assessed as received and all applicants should follow the standard college application procedure (indicating Dr Yunjie Gu as supervisor)
(http://www3.imperial.ac.uk/pgprospectus/howtoapply).

Closing date for applications: Open until filled
Start Date: As soon as possible from September 2018.


Funding Notes

The studentship provides a (tax-free) bursary of £16,553 (Standard RCUK Bursary rate) per annum for up to 3.5 years to cover living expenses, together with the College tuition fees at the UK/EU rate for 3 years. Applications from Overseas will also be considered, but the difference in the Overseas and UK/EU rate will have to be met by the successful applicant.