PhD student, Wei Huang, from the University’s Department of Computer Science has won the ‘most original approach’ category at Siemens Mobility’s Artificial Intelligence Dependability Assessment – Student Challenge (AI-DA-SC) awards.
Siemens Mobility is pioneering transportation, moving people sustainably and seamlessly from the first mile to the last. With the help of AI, they create new intelligent applications that make mobility safer, more comfortable and more efficient. “Machine Learning” (ML) is of great importance in this context. The use of ML techniques for safety decisions is innovative and demanding, presenting complex underlying problems.
Siemens Mobility ran the AI-DA Open Innovation Challenge for students to create an AI model for a simple classification problem with demanding safety requirements. They were asked to develop an AI-model that classifies given datasets and provides stringent and traceable justification for the misclassification probability of the model.
Students from universities worldwide were given the unique opportunity to present their proposals to an expert jury of industrial managers and academic researchers at Siemens Mobility. Three student teams from the University of Liverpool were shortlisted: one from the Department of Computer Science and two from the EPSRC CDT in Distributed Algorithms. The winners were announced at a recent virtual conference.
Wei’s technical solution won the “most original approach” award, winning a total prize fund of €3,000. His project ‘Assessing the Reliability of Machine Learning Model Through Robustness Evaluation and Operational Profiles’ forms the basis of the Defence Science and Technology Laboratory (Dstl) project SOLITUDE.
Wei is currently a third-year PhD student under the supervision of Dr Xiaowei Huang and Dr Xingyu Zhao. His research is focused on testing and verification methods for machine learning models such as deep neural networks and ensemble trees.
Dr Xiaowei Huang, Senior Lecturer, Department of Computer Science said: “I was extremely happy to hear about Wei winning this highly competitive prize. It is certainly a recognition of Wei’s research on AI safety – a promising new direction to make sure AI performs within humans’ expectations.”
Wei added: “I’m very excited and honoured to win this prize. Most importantly, I’m proud that my research product is recognised by the industry. My aim is to keep going on and try my best to contribute to the development of AI safety.”
AI safety is a new research area that has drawn much attention in the past few years. Beyond the solution Wei proposed for the challenge, a few other directions will be explored. If you would like to learn more about Wei’s research you can contact him at W.Huang23@liverpool.ac.uk.
Ilaria Carrara Cagni, Siemens Technology said: “Congratulations to Wei Huang for the successful participation in the Student Challenge ‘AI Dependability Assessment’, where more than 30 teams of brilliant students from 15 countries developed AI Models from the perspective of demanding safety requirements. Artificial Intelligence models are key to solve complex safety issues which are necessary in the process of digitalization and automation of the mobility sector.”
These are excellent examples of how the University’s research in Digital is developing innovative and creative solutions to enable the transformation of society and industry through the generation, communication and processing of data. You can read further information about the Institutional Research Theme here and join the research community group by emailing Anthony Mtitimila, Senior Partnerships & Innovation Manager and Digital Theme Professional Services Lead at Anthony.Mtitimila@liverpool.ac.uk.