Výsledky vyhledávání - "Maurer, Karsten"

Report

Generalized Adversarial Distances to Efficiently Discover Classifier Errors

Autor: Bennette, Walter, Dufek, Sally, Maurer, Karsten, Sisti, Sean, Tusmatov, Bunyod

Given a black-box classification model and an unlabeled evaluation dataset from some application domain, efficient strategies need to be developed to evaluate the model. Random sampling allows a user to estimate metrics like accuracy, precision, and

Externí odkaz: http://arxiv.org/abs/2102.12844

Zobrazit plný text záznamu

Report

Harnessing Adversarial Distances to Discover High-Confidence Errors

Autor: Bennette, Walter, Maurer, Karsten, Sisti, Sean

Given a deep neural network image classification model that we treat as a black box, and an unlabeled evaluation dataset, we develop an efficient strategy by which the classifier can be evaluated. Randomly sampling and labeling instances from an unla

Externí odkaz: http://arxiv.org/abs/2006.16055

Zobrazit plný text záznamu

Report

Facility Locations Utility for Uncovering Classifier Overconfidence

Autor: Maurer, Karsten, Bennette, Walter

Assessing the predictive accuracy of black box classifiers is challenging in the absence of labeled test datasets. In these scenarios we may need to rely on a human oracle to evaluate individual predictions; presenting the challenge to create query a

Externí odkaz: http://arxiv.org/abs/1810.05571

Zobrazit plný text záznamu