WICI Seminar - On ranking merit: applying the page-rank algorithm to the electoral process
FREE to Attend: RSVP with Eventbrite.
![]()
Communities often face the following problem: to collectively achieve a ranking of their members according to some notion of merit using only the opinions of those members and without any pre-existing ranking. If a society wished to appoint only the most meritorious of its members to a particular set of offices --- for instance, the most trustworthy to decision-makers, the most fair to jurors, the most expert to policy-makers--- they would face an instance of this problem. I will consider a solution based on the page-rank algorithm, which is the technological innovation at the root of Google's search engine (and the reason for its dominance over earlier search engines). Roughly speaking, if one member of a community assigns a certain fraction of her vote to another, then her contribution to that person's merit ranking is scaled by the fraction chosen and by her own merit ranking. The rankings are obtained as a self-consistent solution from the complete directed network of voting relations. Although other electoral schemes, such as majority-vote, can be used to try and achieve the same ends, I provide reasons for thinking that the proposed algorithm, and variations thereof, will fare better. Finally, I will discuss the possible failure modes of such a scheme.
Speaker Profile: Robert Spekkens
Robert Spekkens received his B.Sc. in physics and philosophy from McGill University and completed his M.Sc. and Ph.D. in Physics at the University of Toronto. He held a postdoctoral fellowship at Perimeter Institute and an International Royal Society Fellowship at the University of Cambridge. He has been a faculty member at Perimeter Institute since November 2008. His research is focused upon identifying the conceptual innovations that distinguish quantum theories from classical theories and investigating their significance for axiomatization, interpretation, and the implementation of various information-theoretic tasks.










