Perceiving the Social: A Multi-Agent System to Support Human Navigation in Foreign Communities

TL;DR

This paper presents a multi-agent system that models social networks to assist individuals in navigating and engaging with local communities by providing tailored social connection recommendations based on reconstructed social spaces.

Contribution

The novel system reconstructs social spaces from community social network data and uses agent modeling to facilitate human navigation and interaction in social environments.

Findings

Effective social space modeling from network data

Improved navigation guidance in social contexts

Enhanced community engagement strategies

Abstract

This paper describes a system developed to help people explore local communities by providing navigation services in social spaces created by the community members via communication and knowledge sharing. The proposed system utilizes data of a community's social network to reconstruct the social space, which is otherwise not physically perceptible but imaginary, experiential, yet learnable. The social space is modeled with an agent network, where each agent stands for a member of the community and has knowledge about expertise and personal characteristics of some other members. An agent can gather information, using its social "connections", to find community members most suitable to communicate to in a specific situation defined by the system's user. The system then deploys its multimodal interface, which "maps" the social space onto a representation of the relevant physical space, to locate the potential interlocutors and advise the user on an efficient communication strategy for the given community.