Interpretable reinforcement learning for complex manufacturing processes - EPSRC industrial CASE studentship award

A fully funded 4 years PhD studentship (EPSRC industrial CASE award, co-funded by Wayland Additive Limited) within the Department of Automatic Control and Systems Engineering  is available with focus on interpretable reinforcement learning algorithms for the multi-physics control and optimisation of complex manufacturing processes. The manufacturing process under investigation in this project is metallic powder-bed 3D printing (aerospace, biomedical sector).

Machine Learning (ML) research enabled an ‘explosion’ of applications and impact in advanced manufacturing processes in recent years, primarily in proof-of-concept and laboratory-scale demonstrator systems. While the potential use of ML-enabled technologies can be transformative in the sector, it is also recognised that ML alone is not the only answer to realising fully digitalised, advanced automated and autonomous factories. A key bottleneck to adoption is the fundamental understanding of how complex information propagates in ML structures, and how this information can be used to enable advanced automated process control. 

Within this context, fundamental research that focuses on ML interpretability, as well as research that efficiently interfaces ML structures to advanced control theory would be highly influential in the discipline. This project aims to address the following underpinning research question “How can we guarantee robust automated process control using ML structures and algorithms?”.

The successful PhD applicant will join a group of PhD students and post-doctoral researchers led by Professor George Panoutsos. Collectively, the group’s research is focused on interpretable machine learning theory, algorithms and applications, for process modelling, optimisation and control. 

https://www.sheffield.ac.uk/acse/department/people/academic/george-panoutsos

Strong links with the EPSRC MAPP Hub will also provide the PhD student the opportunity to collaborate and interact with a large interdisciplinary centre in the application area.

https://mapp.ac.uk

Entry requirements

The studentship is available to candidates with the equivalent of a first class or upper second-class degree in Control and Systems Engineering, and/or Computer Science and/or Mechanical Engineering. A strong background in Machine Learning and Control theory, with good coding skills, is highly desirable. 

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. More details regarding English Language Qualifications can be found here - https://www.sheffield.ac.uk/postgraduate/english-language