# Dimitris Koukoulopoulos

Université de Montréal

March 14, 2014 from 16:00 to 18:00 (Montreal/Miami time) On location

Colloquium presented by **Dimitris Koukoulopoulos (Université de Montréal)**

When trying to understand extreme phenomena in mathematics, one of the natural things to study is whether the extremizer has any special structure. Indeed, the more information one has on the extremiser, the better one should be ably to analyze the phenomenon under investigation. This approach has been proven very effective when studying the average behaviour of general multiplicative functions. These are complex-valued functions defined over the integers which respect the multiplicative structure of the integers, i.e. f(mn)=f(m)f(n) when m and n are coprime. They are of central importance to number theory as several important questions in number theory can be formulated in terms of the average behaviour of them. Perhaps the most prominent example is the Riemann Hypothesis, which is equivalent to proving that the partial sums of a certain multiplicative function exhibit square-root cancellation. During the recent years, Granville and Soundararajan pioneered a new theory whose goal is to unify and extend the theory of general multiplicative functions. The starting point is a theorem of Halasz which states that if a multiplicative function assumes values inside the unit circle then its partial sums can be large only if it “pretends to be” a very special multiplicative function, the function n^{it} with t fixed. So Halasz’s theorem gives a very elegant description of the extremizers for the problem of maximizing the partial sums of a function. This simple idea, of a one multiplicative function pretending to be another one, turns out to be very potent. Indeed, using it we now have new proofs of famous old theorems, such as the Prime Number Theorem and Linnik’s theorem, concerning the existence of primes in short arithmetic progression. More importantly, the theory of pretentious multiplicative functions has shed light to problems which were previously unattackable, most prominently concerning character sums and the Quantum Unique Ergodicity conjecture. My goal in this talk is to present this new and evolving theory, and some of my contributions to it.

**Address**