Representative and equitable synthetic data: ML algorithms and working practices

Important results from the project

The goal of the project was to facilitate synthetic data use by Swedish industry that fairly represents the diversity found in the original data and/or adjusts for existing biases in the data. This objective was met with research results at Linköping University; software development at project partner Fair AI Data;; recurring workshops on synthetic data with project partner Toyota. The research result is also included in IVA´s 100-list, 2024.

Expected long term effects

The project has led to an increased understanding of the complexity of synthetic data, which will result in new tools, new answers to explainability and transparency requirements, and new policy proposals.

Approach and implementation

The project was implemented according to a proposal, with three partners. The research was based at LiU, the technology development was based at Fair AI Data, and industrial needs were explored at Toyota.