Trust-based Moderation Systems

Submit
cblgh.org

Trust-based Moderation Systems

AI summary of the linked article

This article proposes TrustNet, a trust-based moderation system for distributed chat in which participants choose whom they trust to moderate on their behalf, and that trust can be revoked if it proves misplaced. Trust is assigned on a 0.0 to 1.0 scale within a trust area, and the network is traversed transitively to a path length of 3 to 6 steps. TrustNet ranks peers using Appleseed, a metric proposed in the mid-2000s by Cai-Nicolas Ziegler and Georg Lausen of the University of Freiburg, then clusters the ranking with k-means into three groups and keeps the top two. Directly trusted peers are always included, and a threshold strategy returns an empty list if the source has no medium or high trust assignments.

Recommended by 2 curators
No comments yetWhat caught your eye? Start the discussion.