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

This paper aims to explore the feasibility of neural architecture search(NAS) given only a pre-trained model without using any original training data.This is an important circumstance for privacy protection, bias avoidance, etc.,in real-world scenarios. To achieve this, we start by synthesizing usable datathrough recovering the knowledge from a pre-trained deep neural network. Thenwe use the synthesized data and their predicted soft-labels to guide neuralarchitecture search. We identify that the NAS task requires the synthesizeddata (we target at image domain here) with enough semantics, diversity, and aminimal domain gap from the natural images. For semantics, we propose recursivelabel calibration to produce more informative outputs. For diversity, wepropose a regional update strategy to generate more diverse andsemantically-enriched synthetic data. For minimal domain gap, we use input andfeature-level regularization to mimic the original data distribution in latentspace. We instantiate our proposed framework with three popular NAS algorithms:DARTS, ProxylessNAS and SPOS. Surprisingly, our results demonstrate that thearchitectures discovered by searching with our synthetic data achieve accuracythat is comparable to, or even higher than, architectures discovered bysearching from the original ones, for the first time, deriving the conclusionthat NAS can be done effectively with no need of access to the original orcalled natural data if the synthesis method is well designed.

Data-Free Neural Architecture Search via Recursive Label Calibration