Devreal

DBTB INT Jeremy Howard r

DBTB INT Jeremy Howard r

Recording: DBTB INT Jeremy Howard r

[Music] um Jeremy Howard it's a it's nice to be able to give a keynote that covers both the UX and the life sciences and the more vertical tracks so horizontal tracks because it kind of get a chance to talk a bit about how these things all fit together I think it was great to see how hunters talked talked about some of the similar things to what I talked about from a different perspective yeah it's really nice to bring together some themes like that is cool because people are seeing that they using it first stuff I mean everybody can see now that a lot of the stuff they use on their phone or whatever is clearly being driven by data so i think people get like starting to ask questions like oh how did google photos know that that's a photo of my wife or how did you know how did instagram automatically change my picture in that way or added the google inbox automatically suggests this reply like they all sound seemed kind of magical and kind of awesome and you know i think people want to know how does it work and how can i use how can i do that [Music] our title of my talk kind of gives that away which is that you can't afford to ignore deep learning one of these things which I know a lot of people who are more you know they've seen a lot of trends come and go and because it seems so over fashionable they assume it must be kind of ridiculous you know actually this is one of those rare trends that comes along that actually is every bit as important as the breathless media articles make out it is so I try to give some examples of how deep learning is already changing their day-to-day life and what's coming up over the next few months and hopefully get people to think okay maybe it's worth looking into maybe it's not just another fad I know how to become a data scientist you I don't know how to become a great one if I did I would have done it already that yeah my background is totally non technical and now I like to call myself a data scientist so I guess I've been through that process myself it's all very accessible nowadays there's great courses online on Coursera Udacity there's great books I think the most important thing is to work on interesting problems and let that drive your learning so personally I really enjoy Carol competitions like not necessarily just current ones but you can go back and look at past ones have a go at entering them yourself and then you can after a while take a look and see well what are the winners do what do they do differently how come you didn't come up with that idea and you can learn a lot that way [Music] you