English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Train machine learning models on sensitive user data has raised increasingprivacy concerns in many areas. Federated learning is a popular approach forprivacy protection that collects the local gradient information instead of realdata. One way to achieve a strict privacy guarantee is to apply localdifferential privacy into federated learning. However, previous works do notgive a practical solution due to three issues. First, the noisy data is closeto its original value with high probability, increasing the risk of informationexposure. Second, a large variance is introduced to the estimated average,causing poor accuracy. Last, the privacy budget explodes due to the highdimensionality of weights in deep learning models. In this paper, we proposed anovel design of local differential privacy mechanism for federated learning toaddress the abovementioned issues. It is capable of making the data moredistinct from its original value and introducing lower variance. Moreover, theproposed mechanism bypasses the curse of dimensionality by splitting andshuffling model updates. A series of empirical evaluations on three commonlyused datasets, MNIST, Fashion-MNIST and CIFAR-10, demonstrate that our solutioncan not only achieve superior deep learning performance but also provide astrong privacy guarantee at the same time.

LDP-FL: Practical Private Aggregation in Federated Learning with Local Differential Privacy