Team Influencers

Social networks such as Facebook, Twitter, or Instagram are powerful mediums to carry out marketing and political campaigns. Not only do they enable messages to easily and rapidly reach many users, but it also brings credibility to the messages that are being conveyed; people are more inclined to pay attention to a message or referral coming from a friend or an influencer whom they follow.

There is also an increasing need for many campaigns, like the sharing of public opinions, job or loan announcements, news, etc, to require not only a large reach but also fairness in reaching users across sensitive attributes, such as race, gender, or location, to achieve unbiased viewpoints.

How can you maximize the reach of these campaigns while operating on a limited budget? This is where influence maximization (IM) comes in. It aims to maximize information spread in a network under constraints, and in particular, by selecting the most influential users from which the transmission of a specific message should begin. In short, this is the modeling behind what most of us already understand as influencers in social media.

This project would be the continuation of current work by Puya and a team of researchers (including a previous MIDS student), who have built two analytical solutions that improve state-of-the-art in incorporating IM with fairness constraints. The aim of this summer’s project would be to scope out an approach to incorporate instance topics (the content of what each user shares) into the modeling.