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

Deep learning models are increasingly deployed to edge devices for real-timeapplications. To ensure stable service quality across diverse edgeenvironments, it is highly desirable to generate tailored model architecturesfor different conditions. However, conventional pre-deployment model generationapproaches are not satisfactory due to the difficulty of handling the diversityof edge environments and the demand for edge information. In this paper, wepropose to adapt the model architecture after deployment in the targetenvironment, where the model quality can be precisely measured and private edgedata can be retained. To achieve efficient and effective edge model generation,we introduce a pretraining-assisted on-cloud model elastification method and anedge-friendly on-device architecture search method. Model elastificationgenerates a high-quality search space of model architectures with the guidanceof a developer-specified oracle model. Each subnet in the space is a validmodel with different environment affinity, and each device efficiently findsand maintains the most suitable subnet based on a series of edge-tailoredoptimizations. Extensive experiments on various edge devices demonstrate thatour approach is able to achieve significantly better accuracy-latency tradeoffs(e.g. 46.74\% higher on average accuracy with a 60\% latency budget) thanstrong baselines with minimal overhead (13 GPU hours in the cloud and 2 minuteson the edge server).

AdaptiveNet: Post-deployment Neural Architecture Adaptation for Diverse Edge Environments