Interviewed by: Gao Wei and Saha Arkaprava | SCSE Graduate Students’ Club, March 2021
Dear Prof Zhang, many congratulations for winning the 2021 President’s Young Scientist Award. We are highly honoured to have the opportunity to interview you.
Your main research interest is in Computer Vision. How would you describe Computer Vision to a layman?
I think the term Computer Vision is itself a layman’s term; the vision of a computer. Computer vision is not just like computer graphics that only involves visualizing pictures. It also processes pictures to translate the pictures into semantics. As an example, the eye is not only a visual sensor, but also connected to the brain which understands the visual input. This understanding is the key component of computer vision.
What role does AI have to play in the field of Computer Vision? Can you tell us about some interesting real-world applications which testify to this dependence?
Face identification is one of the most successful applications in computer vision. I also think the video surveillance in China is successful; if you commit a crime, you can never escape. The autopilot in Tesla is great from a technological point of view (though it has ethical and legal issues).
As far as I have heard, Computer Vision is a rapidly evolving field; most computer science researchers flock towards this or similar areas nowadays. What in your opinion is the reason behind this? How do you keep up with this rapid rate of evolution? What are the challenges involved and how do you deal with them?
To be a successful computer vision scientist or successful computer PhD student, you need to be well versed in a lot of topics, including machine learning, data mining and natural language processing. You need to keep up with a lot of rapid progress, which makes it very competitive. As a rule, more competitive topics have greater fortunes and grab more attention from people. From an economic point of view, in any AI company, the salary of PhD graduates is higher in computer vision related positions compared to other positions. From a natural point of view, vision is the major sensor in human beings. Thus, if you want to study artificial intelligence, you must conquer computer vision; otherwise, for instance, robots will be blind.
How did you address research failures during your student life, if any?
I think research is all from your own interest. So, you cannot say that your interest is a failure. No one can evaluate your research as a failure from a utility-driven viewpoint. However, if you encounter any technical failures, then just solve it like a homework in school.
Could you give some suggestions to PhD students (both senior and junior) regarding the way they should pursue their research works?
First, in PhD you are no longer a student per se; you are quite independent, and no one can judge you all the time. This is the mindset change you need while transitioning from your previous degree to a PhD. Second, you have to be careful about your relationship with your supervisor. You should ask for help from your supervisor wisely. Before you seek help from your supervisor, you need to be very well prepared, and break down your questions into manipulable, doable, low level detailed tasks. Third, PhD study is a very memorable experience. You need to cherish all relations, collaborators, peers etc. that you come across. The human resources you engage with and the skills you acquire will be beneficial in your entire life. Also, you need to keep yourself happy.
Thanks Prof Zhang for taking time off your busy schedule for the interview. We have benefited a lot from your advice.