Quadratic Voting (QV) is a voting system designed to allow participants to express the *intensity* of their preferences, not just their direction. Unlike traditional one-person, one-vote systems, QV requires a cost proportional to the square of the number of votes cast for a particular option. This mechanism aims to mitigate the "tyranny of the majority" and ensure that strongly held minority preferences are adequately represented. In the context of the N-Way Self-Evaluating Deliberation (NSED) protocol, QV is innovatively applied as an *activation function* for achieving non-linear consensus among a plurality of distinct expert agents. This application is crucial for constructing emergent composite models, allowing ensembles of smaller, consumer-grade models to collectively match or exceed the performance of much larger state-of-the-art models by aggregating their "votes" or contributions in a nuanced, preference-weighted manner. It enables a more sophisticated aggregation of diverse model outputs within a runtime Mixture-of-Models architecture.
Quadratic Voting is a system where people pay more to cast more votes, with the cost increasing quadratically, allowing them to show how strongly they feel about an option. In advanced AI systems, it can act as an activation function to combine the insights of many smaller models into a powerful, unified decision, improving performance.
QV, Quadratic Voting mechanism, preference-weighted voting
Was this definition helpful?