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 many applications, one works with neural network models trained by someoneelse. For such pretrained models, one may not have access to training data ortest data. Moreover, one may not know details about the model, e.g., thespecifics of the training data, the loss function, the hyperparameter values,etc. Given one or many pretrained models, it is a challenge to say anythingabout the expected performance or quality of the models. Here, we address thischallenge by providing a detailed meta-analysis of hundreds ofpublicly-available pretrained models. We examine norm based capacity controlmetrics as well as power law based metrics from the recently-developed Theoryof Heavy-Tailed Self Regularization. We find that norm based metrics correlatewell with reported test accuracies for well-trained models, but that they oftencannot distinguish well-trained versus poorly-trained models. We also find thatpower law based metrics can do much better -- quantitatively better atdiscriminating among series of well-trained models with a given architecture;and qualitatively better at discriminating well-trained versus poorly-trainedmodels. These methods can be used to identify when a pretrained neural networkhas problems that cannot be detected simply by examining training/testaccuracies.

Predicting trends in the quality of state-of-the-art neural networks without access to training or testing data