Collective Intelligence: Aggregation of Information from Neighbors in a Guessing Game

Area of Science:

• Social Systems and Collective Behavior
• Complex Systems Analysis
• Computational Social Science

Background:

• Complex systems, particularly social systems, exhibit information aggregation and coordinated activity.
• Individual interactions in social systems lead to emergent phenomena like norms, opinions, and scientific progress.
• Examples include collaborative platforms (Wikipedia, Linux) and citizen science initiatives.

Purpose of the Study:

• To investigate the performance of a collective in solving a guessing problem with partial information.
• To analyze the impact of network structure on collective performance.
• To understand individual information aggregation strategies within a network.

Main Methods:

• Conducted an online experiment involving individuals as nodes in an interaction network.
• Each individual possessed partial information and made decisions based on neighbor interactions.
• Measured collective performance by tracking the temporal evolution of accuracy across regular lattices and random networks.

Main Results:

• No statistically significant difference in collective performance was observed between regular lattice and random networks.
• Individual information aggregation behavior was accurately described by a Bayesian model.
• The collective's decision-making strategy was found to be suboptimal compared to simpler decision models.

Conclusions:

• Network topology does not significantly impact collective accuracy in this guessing task.
• Individual Bayesian information aggregation explains decision-making patterns.
• The study establishes a micro-macro link between individual behavior and collective outcomes, revealing suboptimal collective performance.