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

On-device training is essential for user personalisation and privacy. Withthe pervasiveness of IoT devices and microcontroller units (MCUs), this taskbecomes more challenging due to the constrained memory and compute resources,and the limited availability of labelled user data. Nonetheless, prior worksneglect the data scarcity issue, require excessively long training time (e.g. afew hours), or induce substantial accuracy loss (>10%). In this paper, wepropose TinyTrain, an on-device training approach that drastically reducestraining time by selectively updating parts of the model and explicitly copingwith data scarcity. TinyTrain introduces a task-adaptive sparse-update methodthat dynamically selects the layer/channel to update based on a multi-objectivecriterion that jointly captures user data, the memory, and the computecapabilities of the target device, leading to high accuracy on unseen taskswith reduced computation and memory footprint. TinyTrain outperforms vanillafine-tuning of the entire network by 3.6-5.0% in accuracy, while reducing thebackward-pass memory and computation cost by up to 1,098x and 7.68x,respectively. Targeting broadly used real-world edge devices, TinyTrainachieves 9.5x faster and 3.5x more energy-efficient training over status-quoapproaches, and 2.23x smaller memory footprint than SOTA methods, whileremaining within the 1 MB memory envelope of MCU-grade platforms.

TinyTrain: Resource-Aware Task-Adaptive Sparse Training of DNNs at the Data-Scarce Edge