The tool will make data sharing convenient and safe, at a time when organisations need to flexibly utilise all available data to participate in a data-driven and automated attack landscape.
A new tool called ‘DoppelGANger’ employs machine learning techniques to enable companies to exchange data with one another without revealing confidential information.
Developed by researchers at Carnegie Mellon University and technology company IBM, the tool uses utilises generative adversarial networks (GAN), which employ machine learning techniques to synthesise datasets that have the same statistics as the original data. GAN refers to a system made up of neural network models that compete with each other to capture and analyse data.
On the datasets provided, models trained with DoppelGANger-produced synthetic data had up to 43% higher accuracy than models trained with synthetic data from competing tools, the team said in a study titled ‘Using GANs for Sharing Networked Time Series Data: Challenges, Initial Promise, and Open Questions’.
The tool is said to help companies share data with third-party vendors to develop products or services.
The CMU and IBM team says the tool requires no prior knowledge of the dataset and its configurations, as the GANs themselves are able to generalise across different datasets and use cases. This makes the tool highly flexible, the researchers say, and that flexibility is key to data sharing in cybersecurity situations.
The tool will make data sharing convenient and safe, at a time when organisations need to flexibly utilise all available data to participate in a data-driven and automated attack landscape, the team stated. The team aims to expand the tool’s capabilities soon to enable it to handle more complex datasets.