Google DeepMind has announced the release of JAX-Privacy 1.0, a new version of its toolkit for building and auditing differentially private models. Built on JAX, the update is designed to make it easier for researchers and developers to build DP training pipelines that combine state-of-the-art DP algorithms with the scalability provided by JAX.
JAX was introduced in 2020 as a high-performance numerical computing library for large-scale machine learning (ML). Its core features include automatic differentiation, just-in-time compilation, and seamless scaling across multiple accelerators. The surrounding ecosystem includes Flax, which simplifies the implementation of neural network architectures, and Optax, which implements state-of-the-art optimizers.
JAX-Privacy is used to help researchers and developers implement differentially private (DP) algorithms for training deep learning models on large datasets. The original version of JAX-Privacy was introduced in 2022 to enable external researchers to reproduce and validate some of Google DeepMind’s advances on private training. It has since evolved into a hub where research teams across Google integrate their research insights into DP training and auditing algorithms.
The company said JAX-Privacy 1.0 integrates its latest research advances and has been re-designed for modularity. The update is aimed at making private training workflows easier to build while preserving the ability to work at scale.
Source: research.google.
Companies can share verified announcements through Newz9’s international press release submission page.
