Zvika Neeman

Abstract

A principal faces n agents. The principal truly likes only k <= n of these agents, but wants as many as possible of the agents to believe she truly likes them. The principal communicates with the agents by giving them public “likes,” which are observable by all the agents. We characterize the “liking strategy” that robustly maximizes the sender’s objective function. The model is applied to political campaigns, grade inflation, workplace promotions, and liking on social media.

An interesting property of most cryptocurrencies, and Bitcoin in particular, is that the whole history of transactions is accessible to anyone. The objective of the ANR JCJC project BITUNAM (Bitcoin User Network Analysis and Mining) I’m leading is to collect and analyse the transactions on the Bitcoin Blockchain. In this presentation, I’ll introduce what we can and cannot see in that data, introduce the principle of some machine learning techniques we use to better interpret those transactions, and present some examples of usages of transaction flow tracking.

Andrés Moya

Mouez Fodha

Seda Ertac

Silvia Sonderegger