HLF – Day 3

John Hopcroft (https://en.wikipedia.org/wiki/John_Hopcroft)

Prof. Hopcroft gave us the privilege of assisting a crash course on machine learning. It was amazing. He took us by the hand and showed what is machine learning and how relates to artificial intelligence. But at the end he stated that right now artificial intelligence “is pattern recognition in high dimensional space” and that we will need more 40 years, another revolution for artificial intelligence to understand an object function or other important aspects.

 

 

Alexei Efros (https://en.wikipedia.org/wiki/Alexei_A._Efros)

Alexei gave us a very good presentation and showed us the different techniques that exists to work with images, how to deal with them. What we can do now, and what we are not able to do yet. Where we are still failing. And he ended his presentation stating that images should e treated as first class citizens, and that “internet is all about cats”. His work on his vacations photos converting them into Monet’s and other artists are amazing, the Cézanne results are great. And it is also possible to convert a painted landscape into a real image, but from some artists, not a Cézanne right now…

Visual data is messy, is hard to understand, to use it but it is all sitting there…

 

 

Leslei Lamport (https://en.wikipedia.org/wiki/Leslie_Lamport )

The presentation of Lamport was about proofs, the relevance of doing proofs and not just any proof, but what he called 21th century proofs.  He presented his case showing us a proof from “Principia” and showing us that this was what scientist have back on the 17th century and that now we have a lot more. Being more specific in computer science there is even a specific name for failures in proofs, in code, bugs. To put something mathematical into something that a computer can understand is really challenging, and to be able to do that correctly is needed a more concise and good way. And Leslei proposes a technique to do that, even challenges the public to try to create something new and send him.

 

 

 

Jeff Dean – Google Brain Team –

Jeff emphasized that he is here in name of his team, that he will present the work of a lot of people from Google Brain Team. The reality of interconnection of smart cities is at our door, and we should improve and change to make our lives better. And he presented data: Waymo – 3 million self-driven trips (until now) – autonomous vehicle is a reality, is here to stay and with each use it gets better and the costs are becoming lower and in a few years, it will be a reality and a lot of people will be able to use it. There is also, advance in health care – as image recognition of exams. There is a threshold where some doctors send patients to home and to come back in a year and some doctors send to surgery right away, and there is not a correct decision that can be made by doctors. They are now working with  http://tensorflow.org  – machine learning ideas, platform for everyone, make the platform excellent for both research and production use ML classes in Toronto, Berkeley and Yale. They are also working on automated machine learning – learn to learn, and that is not easy to achieve.

 

Out of the blue, and without previous arrangement we did a gathering of women at HLF. And it was great to see a lot of women, and what some accomplished.

 

In the afternoon happened the poster flash presentations where I had 2 minutes to present to the whole HLF. The topics of the flash poster presentation were about everything in computer science and math, and I was trembling more than a tree in a hurricane. The two big screens and the public, were a little bit scarring but everything worked fine at the end. And we finished explaining a bit more about our work on the posters. And I was so thrilled talking about it, that I almost lost the last bus to the Dinner Reception.

At the dinner reception, which theme was “Bavarian Night”, some people from the team danced Bavarian traditional dances and many people come with their traditional clothing and I was dressed as a real “huasa”. The dinner was very good, and seeing everybody dancing trying to learn a bit of Bavarian dance was awesome.