The Story
This biomedical business, which tracks blood and genetic markers, was looking to anonymise its data and generate synthetic datasets with a high similarity to the original data. Chesamel worked with the business to design the most efficient data anonymisation solution and improve datasets using generative AI.
Impact
Our solution helped eliminate data privacy concerns for the organisation, as well as enhancing data quality by fixing errors and outliers in existing data and augmenting missing data. This vastly improved the training performance of the generative AI model.
Methodology
We implemented generative AI models including Variational Auto-Encoder and Generative adversarial network.This allowed us to generate fully anonymised synthetic datasets closely resembling the original data.
/ services provided
