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/

Zendy Plus

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

Zendy Tools

Zendy Open

An open problem in differentially private deep learning is hyperparameteroptimization (HPO). DP-SGD introduces new hyperparameters and complicatesexisting ones, forcing researchers to painstakingly tune hyperparameters withhundreds of trials, which in turn makes it impossible to account for theprivacy cost of HPO without destroying the utility. We propose an adaptive HPOmethod that uses cheap trials (in terms of privacy cost and runtime) toestimate optimal hyperparameters and scales them up. We obtain state-of-the-artperformance on 22 benchmark tasks, across computer vision and natural languageprocessing, across pretraining and finetuning, across architectures and a widerange of $\varepsilon \in [0.01,8.0]$, all while accounting for the privacycost of HPO.

A New Linear Scaling Rule for Private Adaptive Hyperparameter Optimization