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

Machine learning is traditionally studied at the model level: researchersmeasure and improve the accuracy, robustness, bias, efficiency, and otherdimensions of specific models. In practice, the societal impact of machinelearning is determined by the surrounding context of machine learningdeployments. To capture this, we introduce ecosystem-level analysis: ratherthan analyzing a single model, we consider the collection of models that aredeployed in a given context. For example, ecosystem-level analysis in hiringrecognizes that a job candidate's outcomes are not only determined by a singlehiring algorithm or firm but instead by the collective decisions of all thefirms they applied to. Across three modalities (text, images, speech) and 11datasets, we establish a clear trend: deployed machine learning is prone tosystemic failure, meaning some users are exclusively misclassified by allmodels available. Even when individual models improve at the population levelover time, we find these improvements rarely reduce the prevalence of systemicfailure. Instead, the benefits of these improvements predominantly accrue toindividuals who are already correctly classified by other models. In light ofthese trends, we consider medical imaging for dermatology where the costs ofsystemic failure are especially high. While traditional analyses reveal racialperformance disparities for both models and humans, ecosystem-level analysisreveals new forms of racial disparity in model predictions that do not presentin human predictions. These examples demonstrate ecosystem-level analysis hasunique strengths for characterizing the societal impact of machine learning.

Ecosystem-level Analysis of Deployed Machine Learning Reveals Homogeneous Outcomes