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

Recently, Convolutional Neural Network (CNN) has achieved excellentperformance in the classification task. It is widely known that CNN is deemedas a 'black-box', which is hard for understanding the prediction mechanism anddebugging the wrong prediction. Some model debugging and explanation works aredeveloped for solving the above drawbacks. However, those methods focus onexplanation and diagnosing possible causes for model prediction, based on whichthe researchers handle the following optimization of models manually. In thispaper, we propose the first completely automatic model diagnosing and treatingtool, termed as Model Doctor. Based on two discoveries that 1) each category isonly correlated with sparse and specific convolution kernels, and 2)adversarial samples are isolated while normal samples are successive in thefeature space, a simple aggregate gradient constraint is devised foreffectively diagnosing and optimizing CNN classifiers. The aggregate gradientstrategy is a versatile module for mainstream CNN classifiers. Extensiveexperiments demonstrate that the proposed Model Doctor applies to all existingCNN classifiers, and improves the accuracy of $16$ mainstream CNN classifiersby 1%-5%.

Model Doctor: A Simple Gradient Aggregation Strategy for Diagnosing and Treating CNN Classifiers