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

Explanations of time series models are useful for high stakes applicationslike healthcare but have received little attention in machine learningliterature. We propose FIT, a framework that evaluates the importance ofobservations for a multivariate time-series black-box model by quantifying theshift in the predictive distribution over time. FIT defines the importance ofan observation based on its contribution to the distributional shift under aKL-divergence that contrasts the predictive distribution against acounterfactual where the rest of the features are unobserved. We alsodemonstrate the need to control for time-dependent distribution shifts. Wecompare with state-of-the-art baselines on simulated and real-world clinicaldata and demonstrate that our approach is superior in identifying importanttime points and observations throughout the time series.

What went wrong and when? Instance-wise Feature Importance for Time-series Models