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

Ensuring trustworthiness in machine learning (ML) models is amulti-dimensional task. In addition to the traditional notion of predictiveperformance, other notions such as privacy, fairness, robustness todistribution shift, adversarial robustness, interpretability, explainability,and uncertainty quantification are important considerations to evaluate andimprove (if deficient). However, these sub-disciplines or 'pillars' oftrustworthiness have largely developed independently, which has limited us fromunderstanding their interactions in real-world ML pipelines. In this paper,focusing specifically on compositions of functions arising from the differentpillars, we aim to reduce this gap, develop new insights for trustworthy ML,and answer questions such as the following. Does the composition of multiplefairness interventions result in a fairer model compared to a singleintervention? How do bias mitigation algorithms for fairness affect localpost-hoc explanations? Does a defense algorithm for untargeted adversarialattacks continue to be effective when composed with a privacy transformation?Toward this end, we report initial empirical results and new insights from 9different compositions of functions (or pipelines) on 7 real-world datasetsalong two trustworthy dimensions - fairness and explainability. We also reportprogress, and implementation choices, on an extensible composer tool toencourage the combination of functionalities from multiple pillars. To-date,the tool supports bias mitigation algorithms for fairness and post-hocexplainability methods. We hope this line of work encourages the thoughtfulconsideration of multiple pillars when attempting to formulate and resolve atrustworthiness problem.

Function Composition in Trustworthy Machine Learning: Implementation Choices, Insights, and Questions