What should I talk about?

Madhu Sudan (Harvard University)

Presumably I ought to have answered the question before sending this abstract, but let me tell you about some thoughts as I get into it. The talk should presumably be a function of (1) your interests, (2) my knowledge and (3) our common background. And to be honest, I know very little about (1), (2) or (3). And this is true not just about me, but any among us that gets into any conversation. Yet, and quite remarkably, most of us manage to make pretty good use of the limited time available to us, and manage to have fruitful conversations. Given that we study communication and interaction of all sorts in our field, does it not make sense that we should have some explanations of this remarkable everyday phenomena using our models?

In this talk I will give a brief sense of our attempts to explore these ``day-to-day'' communications leading us to the challenges of ``Semantic communication'' (with Brendan Juba), ``Goal-oriented communication'' (with Oded and Brendan) and most recently ``Communication amid uncertainty'' (with a host of collaborators).