Computer Vision News - October 2021

46 Interview Best of MICCAI 2021 something you have to be aware of and nurture. Thinking about your students, when they get their first paper accepted, is your excitement comparable with the excitement you felt in the past when the same thing happened to you? My students are much more successful than me! They work harder, are more motivated, more productive. It’s obviously a function of the competitive environment they’re in. It’s also a function of the resources they have. But it is true that having your paper accepted at something like MICCAI is very valuable in an objective way. It’s going to be an important line in their CV. It can make or break a career sometimes. If you have a few solid papers during your PhD, that might make a big difference in terms of what kinds of career prospects you have versus if you don’t. Sometimes a lot of that is random. I tell my students there’s a lot of noise in the review process. It’s not perfect the way we review papers and decide their fate. It’s important to understand that noise and understand human psychology, how reviews happen and how reviewers respond to certain things. You have to roll the dice multiple times and be persistent because if you declare a loss that means you’ve given up on an idea or a paper and there’s nothing more to do. But the key is to take what you’ve learned and make your paper better and try again. Learn from your mistakes and from the feedback. Also, sometimes acknowledge that the feedback really is just noise, and you can ignore it. It’s important for mentors to point out what’s noise and what’s not. Do you have any wow moments from your students that you would like to share with us? That’s very hard for me! I have many good students and a lot of respect for all of them. I can highlight one though, Meenakshi Khosla, who just graduated and is now going on to be a postdoc at MIT with Nancy Kanwisher in neuroscience. Meenakshi has done some very interesting work where she used deep learning models to characterize how the brain responds to visual and auditory stimuli. There’s a lot of excitement around the idea of attention in machine learning – whether you’re looking at images, text, or language data, if you build models that can incorporate an idea of attention, it will yield better solutions. Meenakshi was building a model where the input was a picture, and the output was how your brain