PhD Studentship Data-Driven (Bayesian) Optimization for Problems with Dynamic Resource Constraints

[Richard Allmendinger]

Candidates are sought for a funded PhD position in Data-Driven (Bayesian) Optimization for Problems with Dynamic Resource Constraints in the Decision and Cognitive Sciences Research Centre, at the Alliance Manchester Business School, The University of Manchester. The position is funded for a period of 4 years by EPSRC and IBM.

The PhD position is associated with an iCASE studentship with IBM and will focus on the development of algorithms and decision support tools for closed-loop problems with dynamic resource constraints. This includes developing data-driven/Bayesian optimization algorithms capable of dealing with closed-loop problems subject to constraints that model the temporal availability of resources needed to evaluate a candidate solution. 

The PhD student will be encouraged to collaborate with peers in the research centre and develop a wide range of skills including project management, mentoring of interns, and presentation skills. They will also be expected to present their work at major international conferences and participate in events linked to the Institute for Data Science and Artificial Intelligence at Manchester. Both the Institute and the Decision and Cognitive Sciences Research Centre are affiliated with the Alan Turing Institute providing access to the resources and network of the Institute through its fellows programme. This includes access to data study groups and interest groups at the Turing.

The student will be supervised by Dr Richard Allmendinger, Dr Manuel Lopez-Ibanez, Dr Jonathan Shapiro, and Prof Joshua Knowles. Dr Matt Benatan, Algorithms Subgroup Lead in Machine Learning and AI at IBM Research UK will act as external supervisor and lead the industrial input into this research. The ICASE studentship also offers the opportunity to be partly based at IBM Research UK.

The team at IBM Research is focused on cutting-edge research in advanced Bayesian modelling, including Bayesian Optimization, and are thus invested in how Bayesian Optimization can be extended to challenging optimization tasks. The project will therefore contribute directly to IBM's ongoing research with the opportunity for influencing future IBM products. 

For instructions on how to apply, please refer to: https://www.jobs.ac.uk/job/BYV319/epsrc-ibm-industrial-case-phd-studentship-in-data-driven-optimization-tuning-bayesian-optimization-for-problems-with-dynamic-resource-constraints