Dr Shufan Yang
- Position: Associate Professor in Computational Intelligence & Optimisation
- Areas of expertise: Computational Intelligence, Optimisation, Edge Acceleration, Image Reconstruction
- Email: S.F.Yang@leeds.ac.uk
- Phone: +44(0)113 343 3492
- Location: 2.44 Mechanical Engineering
Profile
I am an Associate Professor in Computational Intelligence & Optimisation at the Institute of Medical & Biological Engineering, School of Mechanical Engineering, University of Leeds. I was also appointed as an Honorary Associate Professor in Medical Technologies at the Research Department of Orthopaedics and Musculoskeletal Science, University College London. I obtained my PhD degree in Computer Science from the University of Manchester, UK, in 2010 under the supervision of Professor Steve Furber. I was a Royal Academy of Engineering Industrial Fellow from 2022 to 2023 with Codeplay Software Ltd (acquired by Intel Software Inc), Co-I for a Dstl-funded project 'Exploring Decision-Making Processes in Autonomous Systems' in 2015, and PI for two Innovate UK funded Knowledge Transfer Partnerships Projects and Co-I for EU FP7 ‘Towards Ubiquitous Low-power Image Processing Platforms’.
My research has been funded by EU Horizon, Dstl, Innovate UK, EPSRC, the Royal Society of Edinburgh, and the Royal Academy of Engineering. In the past, I have worked on deep reinforcement learning, system-on-chip, and computational neuroscience research areas, publishing over 60 scientific articles in top-tier journals and conferences, including Nature, IEEE Transactions on Medical Imaging, Parallel Computing, etc. I have served on many research award panels for Innovate UK, EPSRC peer review, Royal Academy of Engineering Research Fellow peer review, Canada Foundation for Innovation, and the Dutch Research Council. I have also been involved in multiple STEM public engagement activities with Nuffield Health and currently serve as a council member of the BCS, the Chartered IT Council.
Research interests
My research focuses on using deep neural network algorithms for optimisation in engineering problems, particularly in translational therapeutics for neurodegenerative disease, the analysis of medical and biomedical images for biomarker extraction, and data-driven healthcare utilising edge machine learning acceleration technologies. My current research projects include developing optimisation for mesh-based Monte Carlo photon simulation software on heteogeneous platforms to benefit from parallel processing capabilities, optimised memory management, thread allocation and customised interactive algorithmic optimisations. I am also working on building pathology and physiology modeling for neurodegenerative disease and using optimisation techniques to improve the performance of physics-informed neural networks and covolutional neural networks for spatial data, refining mesh size for computational models and improving algorithms for numerical solutions. I specialise in hardware-software acceleration for artificial neural network imaging reconstruction in ultrasound imaging, photoacoustic computed tomography, optical coherence tomography, and radar-based human activity recognition using micro-Doppler techniques for resource-limited platforms.
I advocate for open science, particularly in developing digital tissue phantoms. One of my ongoing projects can be found on my GitHub page, accessible at WOLVS/standardised-image-reconstruction: The IPASC standardised image reconstruction project.
I am also passionate about open hardware and have participated in competitions for the last decade. My past winning projects include:
- 2023 AI PHD Prize Winner fNIR motion artification detection:2023 Results Gallery - XOHW (openhw.eu);
- 2019 PYNQ White blood cell microscopic image classification:2019 - XOHW (openhw.eu);
- 2016 real-time open access image processing platform:2016 - XOHW (openhw.eu).
Qualifications
- PhD in Computer Science
Professional memberships
- Charted Engineer (CEng)
- Senior IEEE
- MBCS
- FHEA
Student education
- Robotics and Machine Intelligence
- Professional Project