Evaluating Competitive Influence Maximisation Strategies Using an Online Game

Abstract We demonstrate an online game that will help us understand how we can best maximise our influence across a large social network and in the presence of adversaries, for example to gather intelligence data or recruit soft sensors in a conflict zone. Influence maximisation (also referred to as opinion control) is the study of strategically influencing agents on social networks with the aim to align their opinions, behaviours, or choices with certain targets, and has been extensively studied in competitive and non-competitive scenarios, mostly via variants of models based on the seminal independent cascade model [1]. However, these models may not be appropriate in situations in which agents are subject to various sources of social influence, and decisions can be changed over time. The voter model, by allowing agents to repeatedly change their opinion, provides a more accurate description of the underlying opinion dynamics mechanism, as gathering data from soft sensors is often a recurring task. For this reason, and given the criticality of coalition operations, opposing parties can try to strategically influence key individuals to prevent them from contributing information to the alliance.
  • Valerio Restocchi (Southampton)
  • Sebastian Stein (Southampton)
  • Markus Brede (Southampton)
  • Soheil Eshghi (Yale)
Date Sep-2018
Venue 2nd Annual Fall Meeting of the DAIS ITA, 2018