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

Deep Learning (DL) frameworks are now widely used, simplifying the creationof complex models as well as their integration to various applications even tonon DL experts. However, like any other programs, they are prone to bugs. Thispaper deals with the subcategory of bugs named silent bugs: they lead to wrongbehavior but they do not cause system crashes or hangs, nor show an errormessage to the user. Such bugs are even more dangerous in DL applications andframeworks due to the "black-box" and stochastic nature of the systems (the enduser can not understand how the model makes decisions). This paper presents thefirst empirical study of Keras and TensorFlow silent bugs, and their impact onusers' programs. We extracted closed issues related to Keras from theTensorFlow GitHub repository. Out of the 1,168 issues that we gathered, 77 werereproducible silent bugs affecting users' programs. We categorized the bugsbased on the effects on the users' programs and the components where the issuesoccurred, using information from the issue reports. We then derived a threatlevel for each of the issues, based on the impact they had on the users'programs. To assess the relevance of identified categories and the impactscale, we conducted an online survey with 103 DL developers. The participantsgenerally agreed with the significant impact of silent bugs in DL libraries andacknowledged our findings (i.e., categories of silent bugs and the proposedimpact scale). Finally, leveraging our analysis, we provide a set of guidelinesto facilitate safeguarding against such bugs in DL frameworks.

Silent Bugs in Deep Learning Frameworks: An Empirical Study of Keras and TensorFlow