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 real-world machine learning applications, data subsets correspond toespecially critical outcomes: vulnerable cyclist detections are safety-criticalin an autonomous driving task, and "question" sentences might be important to adialogue agent's language understanding for product purposes. While machinelearning models can achieve high quality performance on coarse-grained metricslike F1-score and overall accuracy, they may underperform on criticalsubsets---we define these as slices, the key abstraction in our approach. Toaddress slice-level performance, practitioners often train separate "expert"models on slice subsets or use multi-task hard parameter sharing. We proposeSlice-based Learning, a new programming model in which the slicing function(SF), a programming interface, specifies critical data subsets for which themodel should commit additional capacity. Any model can leverage SFs to learnslice expert representations, which are combined with an attention mechanism tomake slice-aware predictions. We show that our approach maintains aparameter-efficient representation while improving over baselines by up to 19.0F1 on slices and 4.6 F1 overall on datasets spanning language understanding(e.g. SuperGLUE), computer vision, and production-scale industrial systems.

Slice-based Learning: A Programming Model for Residual Learning in Critical Data Slices