School of Engineering, UNIVERSITY OF EDINBURGH

Position ID:UNIVERSITY OF EDINBURGH-School of Engineering-PDRA [#10714, 042263]
Position Title: Postdoctoral Research Associate in Scientific Computing
Position Type:Postdoctoral
Position Location:Edinburgh, Edinburgh, City of EH9 3JL, United Kingdom [map]
Subject Area: Engineering / Computing
Appl Deadline:2018/01/31 (posted 2018/01/06)
Position Description:    

We seek an outstanding researcher for a project addressing randomised algebra algorithms for real-time process analytics. In essence, the aim is to develop statistical computing tools that utilise a priori available information on model structure. This project is funded by EPSRC, and will be based at the School of Engineering of the University of Edinburgh. The successful applicant will conduct research into developing and analysing randomised sampling algorithms for solving partial differential equations on high dimensional models and inverse regression problems to estimate model parameters. The responsibilities include the design of randomisation algorithms for finite element matrix sketching, design of subset selection algorithms for large-scale regression, analysis of the randomisation-induced error to quantify the variance in the estimated quantities, software implementation in Matlab or Python, and publishing peer reviewed articles and presenting work at conferences.

This Research Associate position will contribute to the development of randomised algebra methods for process analytics within the Institute for Digital Communications at the University of Edinburgh. The project is funded by EPSRC. To expedite the data processing within process analytics we suggest replacing the conventional deterministic computations with sample-based estimations involving only a few randomly chosen elements of the model matrices and data vectors. If this sampling is done optimally it could lead to a substantial saving in time and memory requirements, with only a minor degradation in the fidelity of the analytics. The job purpose is to deliver randomisation algorithms for two strategically selected case studies: the first one concerns large systems of linear equations derived by the implementation of the finite element method on the diffusion equation, and the second will be on the large-scale regression problem to estimate the diffusion coefficients. The two cases are well-coupled within the context of process analytics in that typically one would use a numerical scheme to predict the process’s response, while the combination of a finite element model with measured process data naturally leads to a regression inverse problem for diagnostics.

Main Responsibilities

Undertake research related to randomised linear algebra for real-time process analytics. This will entail both independent research work and working in collaboration with the PI and other researchers on randomised algebra for numerical computing. The successful candidate will proactively engage with collaborating researchers to identify fruitful lines of research in this area, with a view to understanding, defining, and quantifying the potential of randomised algebra in randomised algebra for solving partial differential equations and inverse problems. In particular, this involves the design of randomisation algorithms for finite element matrix sketching; design of subset selection algorithms in the context of large-scale regression; analysis of the randomisation-induced error to quantify the variance in the estimated quantities with respect to the optimality of the sampling distributions; and software implementation in Matlab or Python using serial or parallel platforms. (60% of time)

Dissemination of research findings through internal reports, conference proceedings, and journal publications (20% of time)

Liaison with and support of other colleagues on the project, possibly including domestic and international travel and support of meetings at the University of Edinburgh and occasionally to the Alan Turing Institute. (10% of time)

Develop plans for personal research and contribute to wider development of group strategy to identify and agree priorities for current and future work. (5% of time)

Contribute to writing bids to win ongoing research grants. Provide guidance to other staff and students interested in integrating methods/results from this work into their own activity to ensure most effective use, particularly within the University of Edinburgh (5% of time)

Apply at the website: https://www.vacancies.ed.ac.uk/pls/corehrrecruit/erq_jobspec_version_4.jobspec?p_id=042263 or search for vacancy 042263 at https://www.vacancies.ed.ac.uk

This employer is not accepting applications for this position through AcademicJobsOnline.Org. Please see the job description above on how to apply.
Contact: Nick Polydorides, +44 -131-650-2769
Email: email
Postal Mail:
Nick Polydorides
Agile Tomography Group Leader
Institute for Digital Communications
School of Engineering
University of Edinburgh
Rm 2.10, Alexander Graham Bell
King’s Buildings
EH9 3JL, Edinburgh, UK

http://www.homepages.ed.ac.uk/npolydor/
Web Page: https://www.vacancies.ed.ac.uk/pls/corehrrecruit/erq_jobspec_version_4.jobspec?p_id=042263

© 2018 AcademicJobsOnline.Org. All Rights Reserved.