[Paper Review] Minorities report: optimal incentives for collective intelligence.

Richard P. Mann, Dirk Helbing|arXiv (Cornell University)|Nov 11, 2016

Evolutionary Game Theory and Cooperation23 references4 citations

TL;DR

The paper proposes a novel incentive system that rewards accurate minority predictions to sustain diversity and enhance collective intelligence in group decision-making. Using an evolutionary game-theoretic model of collective prediction, it demonstrates that minority-rewarding mechanisms outperform market-based systems by reducing herding, preserving informational diversity, and significantly improving predictive accuracy.

ABSTRACT

Collective intelligence is the ability of a group to perform more effectively than any individual alone. Diversity among group members is a key condition for the emergence of collective intelligence, but maintaining diversity is challenging in the face of social pressure to imitate one's peers. We investigate the role incentives play in maintaining useful diversity through an evolutionary game-theoretic model of collective prediction. We show that market-based incentive systems produce herding effects, reduce information available to the group and suppress collective intelligence. In response, we propose a new incentive scheme that rewards accurate minority predictions, and show that this produces optimal diversity and collective predictive accuracy. We conclude that real-world systems should reward those who have demonstrated accuracy when majority opinion has been in error.

Motivation & Objective

• To investigate how incentive structures influence diversity and collective intelligence in group prediction tasks.
• To identify the limitations of market-based incentives in maintaining informational diversity.
• To design and evaluate a new incentive mechanism that rewards accurate minority predictions.
• To demonstrate that minority-based rewards lead to optimal group predictive performance.

Proposed method

• An evolutionary game-theoretic model is used to simulate repeated collective prediction tasks with heterogeneous agents.
• Agents update their strategies based on payoff feedback, simulating learning and adaptation in social groups.
• The model compares two incentive regimes: market-based rewards (favoring majority choices) and minority-based rewards (favoring accurate minority predictions).
• Key components include payoff functions that reward prediction accuracy and penalize incorrect choices, with differential treatment based on whether a prediction is in the minority or majority.
• The system evaluates long-term outcomes in terms of diversity of predictions, information retention, and collective accuracy.

Experimental results

Research questions

• RQ1 How do market-based incentives affect diversity and collective predictive accuracy in group decision-making?
• RQ2 What are the consequences of herding behavior induced by conventional incentive structures?
• RQ3 Can a reward system that favors accurate minority predictions maintain diversity and improve group performance?
• RQ4 How does minority-based incentive design compare to market-based systems in sustaining collective intelligence?

Key findings

• Market-based incentive systems induce herding behavior, reducing the diversity of predictions and suppressing collective intelligence.
• The minority-rewarding mechanism successfully maintains informational diversity by incentivizing accurate divergent predictions.
• Groups using the minority-incentive scheme achieve significantly higher collective predictive accuracy than those using market-based systems.
• The proposed system outperforms conventional models by aligning incentives with truth-seeking rather than conformity.