A game-theoretic approach to evolving and analysing mechanism design in web-enabled simulation (WES)

Written by Dr Chong Chun Yong, School of Information Technology

How can we build mechanisms into software that help prevent harmful behaviours by making it much harder for users to act in ways that contradict Facebook's community standards? We are keen to explore possible ways the infrastructure and Facebook product code can be improved such that real users have improved experience with the platform whereas users that intend to cause harm will experience friction in their efforts. Our approach will allow testing ways in which Facebook users might harm each other through interactions using the Facebook system, and help make it less likely to happen. The main outcome of this project will be a system (including a theoretical game and its implementation as a tool) that can be used to evolve mechanisms and analyse their impact on the user interaction.

In a networked society, such as the virtually connected users of Facebook, mechanism design and implementation are shaped by the interaction between multiple users. This multi-actor dimension is not new, and theoretical frameworks exist, such as our work in game theory that can form a solid foundation for studying the effect of designed mechanisms by analysing emerging game-theoretic properties during a simulation. Game theory offers a method to develop formal models of such situations and works from the assumption that the outcomes of interactions can be explained from rational choices by individual actors. We aim to evolve mechanisms that prevent actors from reaching a `bad' equilibrium, which results in harmful behaviour.

This project focus on one of the challenges identified by Facebook's Web-Enabled Simulation (WES) research agenda which draws on rich research strands, including machine learning and optimisation, multi-agent technologies, reliability, integrity, privacy, and security as well as traditional simulation, and topics in the user community and emergent behaviour analysis. We are actively recruiting PhD candidates who are interested in the project which is expected to be completed in 2024. Monash University provides an international platform for a researcher to showcase their work which enables me to connect with experts from across campuses to bid for this competitive grant funded by Facebook.