The “Most Popular News” Recommender: Count Amplification and Manipulation Resistance

Abstract

A broad motivation for our research is to build manipulation resistant news recommender systems. There are several algorithms that can be used to generate news recommendations, and the strategies for manipulation resistance are likely specific to the algorithm (or class of algorithm) used. In this paper, we will focus on a common method used on the front page by many media sites of recommending the N most popular articles (e.g., New York Times, BBC, CNN, Wall Street Journal all prominently use this). We show that whereas recommendation of the N most read articles is easily susceptible to manipulation, a probabilistic variant is more robust to common manipulation strategies. Furthermore, for the “N most popular” recommender, probabilistic selection has other desirable properties. Specifically, the (N + 1) th article, which may have just missed making the cut-off, is unduly penalized under common user models. Small differences are easily amplified initially, an observation that can be used by manipulators. Probabilistic selection, on the other hand, creates no such artificial penalty. We use classical results from urn models to derive theoretical results for special cases and study specific properties of the probabilistic recommender.