I am blogging from Lyon where I find myself for the Web Conference 2018. So far, I’ve attended many interesting talks. I enjoyed a lot Jon Kleinberg’s talk on procrastination. George Akerlof noticed that instead of sending a package by mail with one time-effort of cost , he preferred to postpone this every day until the next day. He gave a simple model that explains why procrastination exists by using a multiplicative factor b>1: if he sends the package now, he has to pay a cost of , if he sends it tomorrow the cost will be . Seeing states as nodes in a state-space graph, Akerlof proposed that a procrastinator views the actual cost of the outgoing edges (i.e., possible moves that he can take), at any state multiplied by a factor b and moves myopically/greedily. This elegant model explains many interesting aspects of procrastination (e.g., why start writing two chapters from a book, and then just stop, or why drop a class towards the end of a semester?). While I have to go through this work carefully, it is definitely elegant. I also feel that there is a number of very interesting modeling and algorithmic questions related to this line of work.
Today, I gave a talk on my joint work with Cameron and Chris. The starting point of this work was that one can argue both in favor and against recommending a link between two humans with different opinions on a topic. On the one hand, avoiding this recommendation causes less disagreement. On the other hand, one may argue that if the two humans become connected, and exchange (civilized :-)) arguments, they may develop an understanding for the other person’s views. In other words, they become less polarized. So the key question asked in the paper was at a high level, which graph topology optimizes an objective that takes into account both disagreement and polarization. For the interested ones, the paper can be found here, the slides here, and the project web page here.
Lyon is a beautiful city, with great weather, parks, and other places
but nonetheless I look forward to getting back to Boston already 🙂