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

In recent years, there has been a growing interest in visualizing the losslandscape of neural networks. Linear landscape visualization methods, such asprincipal component analysis, have become widely used as they intuitively helpresearchers study neural networks and their training process. However, theselinear methods suffer from limitations and drawbacks due to their lack offlexibility and low fidelity at representing the high dimensional landscape. Inthis paper, we present a novel auto-encoder-based non-linear landscapevisualization method called Neuro-Visualizer that addresses these shortcomingand provides useful insights about neural network loss landscapes. Todemonstrate its potential, we run experiments on a variety of problems in twoseparate applications of knowledge-guided machine learning (KGML). Our findingsshow that Neuro-Visualizer outperforms other linear and non-linear baselinesand helps corroborate, and sometime challenge, claims proposed by machinelearning community. All code and data used in the experiments of this paper areavailable at an anonymous linkhttps://anonymous.4open.science/r/NeuroVisualizer-FDD6

Neuro-Visualizer: An Auto-encoder-based Loss Landscape Visualization Method