Professor George Panoutsos

MSc, PhD, FHEA

School of Electrical and Electronic Engineering

Head of School

Professor of Computational Intelligence

Professor George Panoutsos
Profile picture of Professor George Panoutsos
g.panoutsos@sheffield.ac.uk
+44 114 222 5130
Amy Johnson Building

Full contact details

Professor George Panoutsos
School of Electrical and Electronic Engineering
Amy Johnson Building
Portobello Street
Sheffield
S1 3JD
Profile

George Panoutsos received his PhD degree in automatic control and systems engineering from the University of Sheffield, Sheffield, U.K, in 2007. He joined the Department of Automatic Control and Systems Engineering (University of Sheffield, UK) as a Lecturer in 2010, and promoted to Professor of Computational Intelligence in 2019.

George has a research grant portfolio of over £3M from the UK EPSRC, Innovate UK, DSTL, EU Horizon 2020 and direct industry funding, as well as over 100 research publications in theoretical as well as applied contributions in the areas of computational intelligence, data-driven modelling, optimisation, control, and decision support systems.

In terms of applied research, the majority of his work is on advanced manufacturing systems, as well as healthcare applications, while also currently exploring research applications in energy and infrastructure.

Research interests

My research focuses on explainable and trustworthy machine learning (ML). Explainability is multifaceted in this context; I work on mathematical and computational methods in Computational Intelligence (CI) that enable enhanced understanding and transparent information use for neural networks, visual and numerical performance measures for many-objective optimisation algorithms, as well as linguistic interpretations of models, and safe control systems. Explainability and trustworthiness are key barriers in using machine learning in a range of critical applications, e.g. in engineering, and healthcare. A multitude of research questions still need to be addressed, for example how neural network - based systems learn and perform when information/data is imperfect, how can we exploit prior knowledge for enhanced learning, and how can we develop performance metrics that will allow us to understand the optimisation of systems at scale.

Towards formulating research questions in machine learning, I often use challenge-driven research e.g. in manufacturing, healthcare, as case studies. This way,  applications drive the research questions, towards maximising impact. I also use explainable machine learning for translational research and to create innovation to address global challenges (e.g. sustainability, energy). The advanced monitoring, optimisation and control of manufacturing processes is such an example, where ML-based methods can be used to reduce material waste, and minimise energy use.

I welcome PhD applications in topics that fall under Computational Intelligence, in particular when these are concerned with explainable machine learning. Examples of recent PhD projects include, physics-guided neural networks, physics-guided generative models, new performance metrics for decomposition-based many-objective optimisation, information theoretic explainability in neural networks, safe reinforcement learning, and linguistic interpretations of Convolutional Neural Networks.

Recent research awards and projects

  • 2019-2021, Innovate UK, 'VULCAN' Data efficient process-part monitoring and control for Electron Beam Melting using Machine Learning (PI, £268k)

  • 2019, UoS ACSE pump-prime fund, A phenotypical knowledge-based image classifier for identification of novel anti-metastasis drugs (PI, £2k)

  • 2019-2021 EPSRC, Using Machine learning to enable feedback controlled manufacture of self-assembled patterned materials (co-I £250k)

  • 2019-2022, Aerospace Technology Institute, DAM, Developing Design for Additive Manufacturing, (co-I, ACSE project £186k)
  • 2019-2022, Aerospace Technology Institute, AIRLIFT, Additive IndustRiaLIsation FuTure Technology, (co-I, ACSE project £140k)
  • 2016-2022 EPSRC MAPP Hub, Future Powder Manufacturing Hub (co-I £10M)
  • 2018-2020 EU H2020, INTEGRADDE (co-I £12.7M)
  • 2017-2019 Innovate UK, MIRIAM - Machine Intelligence for Radically Improved Additive Manufacturing (co-I £666k)
  • 2016 -2018 Innovate UK, TACDAM, Tailorable & Adaptive Connected Digital Additive Manufacturing (academic PI £1M )
  • 2018-2020 TWI Ltd, Phased array NDT in Stir Welding: Interpretable machine learning and process monitoring (PI £42k)
  • 2015-2017 EU H2020, Factories of The Future - 01: Process Optimisation of Manufacturing Assets, COMBILASER, (co-I and academic lead, £3.48M)
  • 2014-2016 TSB, Sustained Process Excellence through Embedding of Analytics and Knowledge Management into Process Chains, Academic (PI, total project cost £441k)
  • 2013-2014 METRC Innovation Award, Online and real-time condition monitoring of Friction Stir Welding, (PI £10k)
  • 2012 EPSRC/Sheffield University, Model-based performance evaluation for critical manufacturing processes (PI £61k)
  • 2012-2014 TWI Ltd. Yorkshire, UK, Automated Systems for Intelligent Stir Tracking and Optimisation (PI £29k)
  • 2010-2013 TWI Ltd. Cambridge, UK, Multiscale model-based search for optimal Process Operating Windows in Friction Stir Welding (PI £6k)
Publications

Journal articles

Chapters

Conference proceedings papers

Patents

  • Mahfouf M, Linkens DA, Panoutsos G & Chen MY () Neuro-Fuzzy Systems. WO/2006/103451 Appl. 01 Jan 1970. RIS download Bibtex download

Preprints

Grants

Current Grants

Previous Grants

  • AIRLIFT: Additive IndustrRiaLIsation for Future Technology), Innovate UK, 01/12/2018 - 30/11/2023, £545,174, as Co-PI
  • Machine Learning digital twin for defect-free additive manufacturing, Research England, 01/02/2022 - 30/06/2022, £29,551, as Co-PI
  • Materials 4.0, RCUK, 01/01/2022 - 31/03/2022, £54,647, as PI
  • CMAC Feasibility Study, RCUK, 01/10/2021 - 30/09/2022, £59,982, as PI
  • VULCAN, Innovate UK, 01/11/2019 - 31/01/2022, £334,563, as PI
  • Using Machine learning to enable feedback controlled manufacture of self-assembled patterned materials, RCUK, 30/09/2019 - 28/03/2022, £252,938, as Co-PI
  • TACDAM: Tailorable & Adaptive Connected Digital Additive Manufacturing, RCUK, 01/01/2017 - 31/12/2018, £221,611, as PI
  • MIRIAM: Machine Intelligence for Radically Improved Additive Manufacturing, Innovate UK, 01/10/2017 - 31/03/2019, £261,312, as Co-PI
  • Integrated machine-part multi-objective optimisation for powder manufacturing, RCUK, 01/11/2016 - 3103/2017, £40,000, as PI
Teaching activities
  • ACS6402, Industry Training Programme in Advanced Manufacturing (module leader)
  • ACS6403, Industry Training Programme in Computational Intelligence (module leader)

Links

Computational Intelligence Systems Laboratory