The Human

M

y name is Nathan, I am to some extent a human being, I live in the most modern time in history, and like many other human beings in this time, I earn my meals by arranging pixels.

I am a theoretical computer scientist. Theoretical computer science is something that can happen when you don’t understand what “information” is, and it bothers you so much that you can’t sleep, and you have no option except to try to undestand what “information” is. This is achieved by arranging pixels into “PDFs”.

I was indoctrinated in the ways of theoretical computer science by Eric Blais at the University of Waterloo. Now I am a postdoc in Mika Göös’ group at EPFL.

Contact: nathaniel.harms (at) epfl.ch

The Science

I

like to study algorithms that are as ignorant as possible, so that I know how they feel. I spend most of my time thinking about two things: property testing vs. learning, and randomized communication.

Property testing vs. learning: The goal is to understand which properties (i.e. sets of functions) can be tested more efficiently than they can be learned. I focus on the distribution-free sample-based model, because in this model, learning is well understood while property testing is not.

Randomized communication: The goal is to understand the power of randomness in communication. I am trying to understand the most extreme examples: the problems where the number of bits of communication required to solve the problem is constant, i.e. independent of the problem size.

The Nonsense

L

ong ago, in the dark and distant past, I was a fine arts student who dropped out after my work was criticized as “too analytical”. Now, on the right occasion, when theorems are evasive, and the stars and the chakras align, under the clearest of skies, when the campfire burns down to embers and its smoke tickles the noses of angels, my pen puts pigment on the paper in a particular pleasant pattern, which some might even say, could crudely be considered “artistic”.