Computer Vision News - October 2018

Daily Wednesday I think part of it is that we don’t have enough people teaching them because the demand has skyrocketed. The number of professors and the number of computer vision classes has not kept pace and at some point, even started to go down. I think this is kind of a longer-term worry that we need to figure out and address somehow. I don’t know how, but I think this is something we should be thinking and worrying about. Can you tell me a funny story about a time when computer vision did not work? When I was starting, there was a saying in vision which was a joke, but every joke has a part of it that is a joke and the other part is the truth. The saying was that if you have a single result image in your paper, it’s a hack. If you have two images, that’s an algorithm. If you have three images, that’s a framework. [ both laugh ] You have already dedicated several decades to science and research. What keeps you thirsty for more knowledge? I think it’s very humbling, because the more you know, the more you realise that you didn’t actually know anything. It’s also that our field has been moving in such crazy ways with so many twists and turns. Once you learn something and you think you understand it, the field moves in a completely different direction and again, you have to relearn everything that you know. That drives the curiosity and the interest. It’s not like we figured everything out and now it’s only the small things. A lot of the big issues are still wide open, and we have no idea how to tackle them. I think that is exciting. That reminds me an answer I received during an interview : “ If you don’t have a crisis during your PhD, you’re not doing it right .” For those students who feel like they are close to a crisis, what would you advise them? I think there are some things that are necessary to be a good researcher. Failure is necessary. My friend Jianbo Shi likes to say that a student needs to suffer. You can’t experience the sweetness of victory without the agony of defeat beforehand. It’s all about the derivative. The springtime flowers smell so sweet because there was a long winter through which you had to suffer first. If things are going too well, it might be that you have picked too easy a problem. It’s very important to challenge yourself. In research, most of the things you are going to work on are not going to work out. This is just how it is. Research is 90% hopeless defeat and 10% elation. Also, in between, the stubbornness of the scientist. Yeah, I guess it’s the stubbornness. It’s also just you really like that 10%, and you wish to have that again, because that does feel good. Victory does taste good. On one hand, there is a need for suffering. You will have to suffer. On the other hand, it is also important to realise that maybe you have suffered Alyosha Efros 14 “Failure is necessary” “… a student needs to suffer”