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

Developers need to perform adequate testing to ensure the quality ofAutomatic Speech Recognition (ASR) systems. However, manually collectingrequired test cases is tedious and time-consuming. Our recent work proposesCrossASR, a differential testing method for ASR systems. This method firstutilizes Text-to-Speech (TTS) to generate audios from texts automatically andthen feed these audios into different ASR systems for cross-referencing touncover failed test cases. It also leverages a failure estimator to findfailing test cases more efficiently. Such a method is inherentlyself-improvable: the performance can increase by leveraging more advanced TTSand ASR systems. So in this accompanying tool demo paper, we devote moreengineering and propose CrossASR++, an easy-to-use ASR testing tool that can beconveniently extended to incorporate different TTS and ASR systems, and failureestimators. We also make CrossASR++ chunk texts from a given corpus dynamicallyand enable the estimator to work in a more effective and flexible way. Wedemonstrate that the new features can help CrossASR++ discover more failed testcases. Using the same TTS and ASR systems, CrossASR++ can uncover 26.2% morefailed test cases for 4 ASRs than the original tool. Moreover, by simply addingone more ASR for cross-referencing, we can increase the number of failed testcases uncovered for each of the 4 ASR systems by 25.07%, 39.63%, 20.9\% and8.17% respectively. We also extend CrossASR++ with 5 additional failureestimators. Compared to worst estimator, the best one can discover 10.41% morefailed test cases within the same amount of time.

CrossASR++: A Modular Differential Testing Framework for Automatic Speech Recognition