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

Backpropagation (BP) has been a successful optimization technique for deeplearning models. However, its limitations, such as backward- andupdate-locking, and its biological implausibility, hinder the concurrentupdating of layers and do not mimic the local learning processes observed inthe human brain. To address these issues, recent research has suggested usinglocal error signals to asynchronously train network blocks. However, thisapproach often involves extensive trial-and-error iterations to determine thebest configuration for local training. This includes decisions on how todecouple network blocks and which auxiliary networks to use for each block. Inour work, we introduce a novel BP-free approach: a block-wise BP-free (BWBPF)neural network that leverages local error signals to optimize distinctsub-neural networks separately, where the global loss is only responsible forupdating the output layer. The local error signals used in the BP-free modelcan be computed in parallel, enabling a potential speed-up in the weight updateprocess through parallel implementation. Our experimental results consistentlyshow that this approach can identify transferable decoupled architectures forVGG and ResNet variations, outperforming models trained with end-to-endbackpropagation and other state-of-the-art block-wise learning techniques ondatasets such as CIFAR-10 and Tiny-ImageNet. The code is released athttps://github.com/Belis0811/BWBPF.

Unlocking Deep Learning: A BP-Free Approach for Parallel Block-Wise Training of Neural Networks