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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Interpretable and explainable machine learning has seen a recent surge ofinterest. We focus on safety as a key motivation behind the surge and make therelationship between interpretability and safety more quantitative. Towardassessing safety, we introduce the concept of maximum deviation via anoptimization problem to find the largest deviation of a supervised learningmodel from a reference model regarded as safe. We then show howinterpretability facilitates this safety assessment. For models includingdecision trees, generalized linear and additive models, the maximum deviationcan be computed exactly and efficiently. For tree ensembles, which are notregarded as interpretable, discrete optimization techniques can still provideinformative bounds. For a broader class of piecewise Lipschitz functions, weleverage the multi-armed bandit literature to show that interpretabilityproduces tighter (regret) bounds on the maximum deviation. We present casestudies, including one on mortgage approval, to illustrate our methods and theinsights about models that may be obtained from deviation maximization.

On the Safety of Interpretable Machine Learning: A Maximum Deviation Approach