# Frustratingly Easy Truth Discovery

## TL;DR

This paper introduces a simple heuristic for truth discovery that estimates worker competence based on average proximity, effectively distinguishing high and low quality sources and improving aggregation accuracy in crowdsourcing.

## Contribution

It proposes a straightforward, effective method for estimating worker competence and demonstrates its theoretical soundness and practical advantages over existing algorithms.

## Key findings

- The heuristic accurately estimates worker competence across various domains.
- Weighted aggregation based on proximity significantly outperforms unweighted methods.
- Under Gaussian noise, the estimate aligns with the maximum likelihood solution.

## Abstract

Truth discovery is a general name for a broad range of statistical methods aimed to extract the correct answers to questions, based on multiple answers coming from noisy sources. For example, workers in a crowdsourcing platform. In this paper, we consider an extremely simple heuristic for estimating workers' competence using average proximity to other workers. We prove that this estimates well the actual competence level and enables separating high and low quality workers in a wide spectrum of domains and statistical models. Under Gaussian noise, this simple estimate is the unique solution to the MLE with a constant regularization factor. Finally, weighing workers according to their average proximity in a crowdsourcing setting, results in substantial improvement over unweighted aggregation and other truth discovery algorithms in practice.