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/

Zendy Plus

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

Zendy Tools

Zendy Open

ML is being deployed in complex, real-world scenarios where errors haveimpactful consequences. In these systems, thorough testing of the ML pipelinesis critical. A key component in ML deployment pipelines is the curation oflabeled training data. Common practice in the ML literature assumes that labelsare the ground truth. However, in our experience in a large autonomous vehicledevelopment center, we have found that vendors can often provide erroneouslabels, which can lead to downstream safety risks in trained models.  To address these issues, we propose a new abstraction, learned observationassertions, and implement it in a system called Fixy. Fixy leverages existingorganizational resources, such as existing (possibly noisy) labeled datasets orpreviously trained ML models, to learn a probabilistic model for finding errorsin human- or model-generated labels. Given user-provided features and theseexisting resources, Fixy learns feature distributions that specify likely andunlikely values (e.g., that a speed of 30mph is likely but 300mph is unlikely).It then uses these feature distributions to score labels for potential errors.We show that FIxy can automatically rank potential errors in real datasets withup to 2$\times$ higher precision compared to recent work on model assertionsand standard techniques such as uncertainty sampling.

Finding Label and Model Errors in Perception Data With Learned Observation Assertions