Participation in this workshop requires additional registration and has limited capacity. If you are interested, please register here
Privacy on Beam is an easy to use & end-to-end differential privacy solution for Apache Beam.
Differential Privacy is a mathematical concept for anonymization and protecting user privacy that has been gaining more and more traction in research and in the industry (for example, US Census is using differential privacy for their 2020 census). However, it is difficult to implement in practice with many pitfalls. There are many privacy-critical steps such as noise addition, partition selection and contribution bounding; and if done incorrectly, could lead to privacy risks. Privacy on Beam is an out-of-the-box differential privacy solution in the sense that it takes care of all the necessary steps for differential privacy without requiring any differential privacy expertise. It is meant to be used by developers, data scientists, differential privacy experts, and more. It is also designed in a way that is compatible & similar with the core Apache Beam SDK, so that developers can convert their pipelines to use differential privacy seamlessly.
We will give a brief introduction into differential privacy and why it is useful and talk about Privacy on Beam. We’ll also have a tutorial/codelab (covering https://codelabs.developers.google.com/codelabs/privacy-on-beam/) to show how to use Privacy on Beam.