After 16 months without lecturing to an audience in my same location, I gave yesterday two lectures at the Technion in front of a live audience (and some additional audience in remote locations). The main lecture was in COMSOC 2021, an international conference on computational social choice, and earlier I gave a guest lecture in Roy Meshulam’s class about simple polytopes. I also met many friends.
Reshef Meir who organized (with Bill Zwicker) COMSOC 2021 wrote:
Hi all,today was beyond expectations – the first feeling of a real actual conference after almost a year and a half! We had about 40 people attending, viewing posters, and listening to talks. I truly hope this will return to be a common scene and that we can all meet face to face soon.
In my COMSOC lecture I talked about some earlier ideas and results in my work on social choice, starting with my paper with Ariel Rubinstein and Rani Spiegler on rationalizing individual choice by multiple rationals, and my subsequent attempt to use learnability as a tool for understanding choices of economic agents. This led to interesting questions on social choice that are discussed in this 2009 post.
In Roy’s course I explained -vectors of polytopes and the Dehn-Sommerville relations based on counting outdegrees of the graph of the polytope when we direct its edges based on a generic abstract objective function. I moved on to present a proof of Blind-Mani’s theorem that the graph of the polytope determines the full combinatorics. This proof is probably the one proof I presented the most and it is given in this 2009 post.
In my COMSOC lecture I described how to fill the two question marks in the table above.