91ÊÓƵ

Zachary Kilpatrick, Department of Applied Mathematics, University of Colorado Boulder

Heterogeneity improves speed and accuracy in social networks

How does temporally structured private and social information shape collective decisions? ToÌýaddress this question weÌýconsider a clique of rational agents who independently accumulateÌýprivate evidence that triggers a decision uponÌýreaching a threshold. Agents’ individual beliefs are given by drift-diffusion equations that receive pulsatile inputs from other deciding agents, according to a normative model derived using Bayesian sequential updating. When seen by the wholeÌýnetwork, the first agent’s choice initiates a wave of new decisions; laterÌýdecisions have less impact. The time and accuracy of the group decision is determined by the extremal statistics of a first passage time problem. In heterogeneous networks, first decisions are made quickly by impulsive individuals who need littleÌýevidence to make aÌýchoice, but, even when wrong, can reveal the correct options to nearly everyoneÌýelse.ÌýWe conclude that groups comprised of diverse individuals can make more efficient decisionsÌýthan homogenous ones. This is joint work with Bhargav Karamched (Florida State), KreÅ¡imir Josić (U Houston), Megan Stickler (U Houston), Benjamin Lindner (Humboldt U Berlin), and Will Ott (U Houston).