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

The recently proposed Activation Relaxation (AR) algorithm provides a simpleand robust approach for approximating the backpropagation of error algorithmusing only local learning rules. Unlike competing schemes, it converges to theexact backpropagation gradients, and utilises only a single type ofcomputational unit and a single backwards relaxation phase. We have previouslyshown that the algorithm can be further simplified and made more biologicallyplausible by (i) introducing a learnable set of backwards weights, whichovercomes the weight-transport problem, and (ii) avoiding the computation ofnonlinear derivatives at each neuron. However, tthe efficacy of thesesimplifications has, so far, only been tested on simple multi-layer-perceptron(MLP) networks. Here, we show that these simplifications still maintainperformance using more complex CNN architectures and challenging datasets,which have proven difficult for other biologically-plausible schemes to scaleto. We also investigate whether another biologically implausible assumption ofthe original AR algorithm -- the frozen feedforward pass -- can be relaxedwithout damaging performance.

Investigating the Scalability and Biological Plausibility of the Activation Relaxation Algorithm