Select Page

Hujun Yin

Professor in Artificial Intelligence


What are your research interests?

I am a professor of artificial intelligence at the University of Manchester. My research areas include neural networks, machine learning, deep learning, image processing and vision systems with strong emphasis on engineering and interdisciplinary applications. We are in the era of AI, a new methodology for problem solving. My aim is to realise maximum and positive impact of AI in growing number of fields. My core activities span from deep learning mechanisms, image recognition, object detection, temporal signal or time series modelling and prediction. Activities and projects include explainable machine/deep learning by the use of manifolds to model the variations, medical image analysis, machine learning based surrogate models in fluid dynamics and heat transfer modelling, and AI-based vision systems for industrial inspections. My collaboration with other disciplines has grown considerably as I see the importance of machine learning and image analysis to modern science and engineering. 

What is the focus of your current research?

Over the last few years, we have developed cutting-edge AI-based vision solutions to object detection in industrial inspections and automation. For example, we have built a prototype system for screening waste streams for lithium ion batteries and quantifying waste compositions in recycling. Such work has now expended to defects detection and quality control in plate glass and building industries. We have also been collaborating with colleagues in chemical engineering, medicine and agriculture in various image analysis and classification work. Current on-going projects include machine learning for early detection of cassava brown streak virus from leaf multispectral images, prostate cancer diagnosis from infrared spectroscopy and H&E stained microscopic images.

What are some projects or breakthroughs you wish to highlight?

We are working towards an efficient AI-enhanced vision system for a wide range of industries so to make industrial inspection and recycling operations more efficient, safer and more automated. One of our prototype systems was a finalist in the 2022 Digital City Awards.

Another recent breakthrough was the use of multispectral imaging and machine learning for detecting cassava brown streak viral infection. The collaborative work in past three years demonstrated feasibility of early detection, significantly earlier than the expensive PCR test. The work, involving NC-State university and Rutgers University in USA, Rothamsted Research, IITA in Tanzania, and Manchester, has recently been awarded $3m (NSF) and £770k (BBSRC) for developing lost-cost devices for the use in the field. Such devices would enable farmers in Africa to root out infected plants and improve crop management and yield.

What memberships and awards do you hold/have you held in the past?

I have been a Turing Fellow since 2018 and hold a senior membership of the IEEE. I am a member of EPRSC College and a member of UKRI Talent Panel College.

What is the biggest challenge in Data Science and AI right now?

How to make the AI development sustainable, accountable, transparent and safer and how make the AI resources accessible fairly across all industries, not just few larger companies, is a huge challenge to the society. 

What real world challenges do you see Data Science and AI meeting in the next 25 years?

Although LLMs have had great success and flourished recently, there are numerous development tasks, esp. interdisciplinary, that require considerable effort from both AI community and the domain experts.


Find out more about Hujun’s research at Research Explorer.