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 Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Currently, many theoretical as well as practically relevant questions towardsthe transferability and robustness of Convolutional Neural Networks (CNNs)remain unsolved. While ongoing research efforts are engaging these problemsfrom various angles, in most computer vision related cases these approaches canbe generalized to investigations of the effects of distribution shifts in imagedata. In this context, we propose to study the shifts in the learned weights oftrained CNN models. Here we focus on the properties of the distributions ofdominantly used 3x3 convolution filter kernels. We collected and publiclyprovide a dataset with over 1.4 billion filters from hundreds of trained CNNs,using a wide range of datasets, architectures, and vision tasks. In a first usecase of the proposed dataset, we can show highly relevant properties of manypublicly available pre-trained models for practical applications: I) We analyzedistribution shifts (or the lack thereof) between trained filters alongdifferent axes of meta-parameters, like visual category of the dataset, task,architecture, or layer depth. Based on these results, we conclude that modelpre-training can succeed on arbitrary datasets if they meet size and varianceconditions. II) We show that many pre-trained models contain degeneratedfilters which make them less robust and less suitable for fine-tuning on targetapplications.  Data & Project website: https://github.com/paulgavrikov/cnn-filter-db

CNN Filter DB: An Empirical Investigation of Trained Convolutional Filters