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

We present a new direction for increasing the interpretability of deep neuralnetworks (DNNs) by promoting weight-input alignment during training. For this,we propose to replace the linear transforms in DNNs by our B-cos transform. Aswe show, a sequence (network) of such transforms induces a single lineartransform that faithfully summarises the full model computations. Moreover, theB-cos transform introduces alignment pressure on the weights duringoptimisation. As a result, those induced linear transforms become highlyinterpretable and align with task-relevant features. Importantly, the B-costransform is designed to be compatible with existing architectures and we showthat it can easily be integrated into common models such as VGGs, ResNets,InceptionNets, and DenseNets, whilst maintaining similar performance onImageNet. The resulting explanations are of high visual quality and performwell under quantitative metrics for interpretability. Code available athttps://www.github.com/moboehle/B-cos.

B-cos Networks: Alignment is All We Need for Interpretability