Last summer I gave a mini-course on the Sum of Squares algorithm in the Swedish Summer School of Computer Science.
It was a great experience – the venue was Djurönäset – a hotel in the beautiful Stockholm archipelgo with stunning views and great food. It was organized very smoothly by Jakob Nordström, Per Austrin, and Johan Håstad, andthe students were bright and enthusiastic. I could see them collaborating on homework problems well into the night, and I’m sure that beyond learning on theoretical computer science, they made great new connections.
In fact, I believe the only downside to that summer school was my own course – this was the first time I taught this material and so it felt very much like a beta version. (On the other hand, Ryan O’Donnell gave a fantastic lecture series on analysis of Boolean function, so the students didn’t suffer too much.) Luckily, this year the summer school features two absolute top experts on coding theory, Venkat Guruswami and Sergey Yekhanin, who will give courses on developments that are not yet as widely known in the theoretical computer science community as they should be, but have had wide reaching impact. Polar codes are considered one of the great recent breakthroughs in coding theory as they provide in some sense the first explicit codes that achieve Shannon’s non-constructive bounds for the capacity of the binary symmetric channel. Codes with local decoding are a prime example of the surprising interplay between theory and practice. Originally arising from probabilistically checkable proofs, they have recently found very significant industrial applications, saving companies 100’s of millions of dollars.
I hear that there are still a few slots left. If you are a Ph.D student, Masters student, or postdoc who want to know more about coding theory, I can hardly think of a more pleasant and beneficial way to do so. If you are interested, please apply as soon as possible at http://s3cs.csc.kth.se/application.php . For any questions, the organizers can be contacted at email@example.com.