| ContentODPDescription
|
We contribute towards improving the interpretation of the biases affecting ML systems while also accounting for them. We see this ODP as a building block to produce reliable and bias-aware documentation of machine learning pipelines
|
| CoversRequirement
|
Who were the participants in the Bias even … Who were the participants in the Bias event?
Who was involved as bias recipient in a bias event?
Who was involved as bias source in a bias event?
What type of bias is documented?
What bias detection method was used?
What is the implication of the bias event documented?
Who are the bias recipient agents that do not have a bias source agent?
What are the bias events for which the bias recipient role is unknown?
What are the bias events for which the bias source role is unknown? for which the bias source role is unknown?
|
| GraphicallyRepresentedBy
|
BiasEventPattern.jpg +
|
| HasIntent
|
An Ontology Design Pattern to Document Bias Events in Machine Learning Pipelines
|
| Modification dateThis property is a special property in this wiki.
|
21 July 2022 15:35:16 +
|
| Name
|
Bias Event Pattern +
|
| OWLBuildingBlock
|
https://labs.tib.eu/sdm/data/BiasEvent.owl +
|
| RelatedCP
|
Https://github.com/kracr/bias-ontology +
|
| Scenario
|
While the Bias Event ODP presented in this work is intended to be used with any machine learning pipeline,
our use case is motivated by the surge of fake news classification systems.
|
| SubmittedBy
|
MayraRusso +
|
| Categories |
ProposedContentOP +
|
