Synthetic Data Sandpit

The Sandpit will offer participants the opportunity to network and build small teams around ongoing challenges related to synthetic data. Following the Sandpit, individuals and groups will have the opportunity to bid for pilot research support to be undertaken by Newcastle’s Research Software Engineering (RSE) team.

Synthetic data are plausible data that are algorithmically generated in order to either replace or to augment existing datasets. The growing interest in generative algorithms is generally motivated by the need to preserve the confidentiality of real data, typically in sensitive application domains (e.g., healthcare), or by the need to augment training sets, typically used in Deep Learning, which are unbalanced or biased, i.e., where data points are scarce in specific regions of interest in the data space.

The primary aim of this sandpit event is to elicit ideas for short, implementation-focused projects in the area of synthetic data generation, validation, and practical deployment for specific applications. The event will also provide a focal point for the growing synthetic data community within the Turing Network.

We anticipate funding up to 5 projects as an outcome of this Sandpit. Typical project awards will be £5,000–£10,000 and funding can be used for Newcastle RSE time or cloud computing costs.

Host or Publisher Newcastle University