Think twice, code once
Zeju Li presenting his winning project at the University Challenge Award Ceremony at the British Museum in London on 12 December 2018.
A vision of AI and some hands-on advice from last year’s winner of Huawei’s UK University Challenge
Zeju Li, currently a first-year PhD student at Imperial College London, scooped the first prize in Huawei’s 2018 UK University Challenge, focusing on computational photography de-noising.
In photography, noise refers to visual distortion. This occurs, for instance, when photos are taken under extreme conditions such as dark night or at high speed. AI is an effective means of removing noise to restore original images.
Zeju achieved the best results by training a deep neural network to formulate this recovery task.
We asked him about his experience of participating in the contest and perspective on how AI will impact our future.
What do you take away from the University Challenge?
I am very grateful to Huawei for providing this wonderful opportunity. I practised and enhanced my engineering abilities during the competition. I had to implement certain engineering processes, such as building dataflow and backend design, and to learn state-of-the-art algorithms in the related CV community.
What are your plans for the coming months and years?
I am currently a first-year PhD student. I will spend more than two years focusing on my research topic, which is medical image computing.
Why did you get into coding?
Nowadays, information technology provides many conveniences, and I think it is useful for everyone to have some basic coding abilities to facilitate their daily work and life.
In your opinion, how can AI change the world? How can we make sure it’s a force for good, triggering positive changes?
In recent years, AI has made considerable changes to our world, especially in the field of computer science. Five years ago, neural networks were rarely mentioned in academic papers; today, it is hard for researchers to get a paper published without covering deep learning methods. I do not think technology can be a force of evil, but people can. We should encourage the development of AI and create a code of conduct for people using or working on it.
In which areas do you see AI having the greatest impact on people’s daily lives in the near future?
From my perspective, AI has ground-breaking capabilities in helping us perceive and recognise the real world. It has already succeeded in helping people carry out easy tasks such as improving image quality and recommending information. If AI develops smoothly over the coming years, I think it can deliver automatic driving and healthcare applications.
What would be your advice to tech students who are thinking about getting involved in the challenge?
Think twice, code once. Don’t start “doing something” before you fully understand the problem and have acquired sufficient background knowledge.
Highlights form the University Challenge were showcased at the DigitALL 2019 New Year Reception on 7 February 2019.