… the information about the features causing bias, we propose a preprocessing approach … machinelearning by reducing the bias caused by several features such as age, race and gender…
… Genderbias in machinelearning occurs when supervised models predict based on spurious societal correlations in their training data. This may result in harmful behaviour when it …
A Farkas, R Németh - Social Sciences & Humanities Open, 2022 - Elsevier
… genderbias in machine translation through a case study of Google Translate. We think that by investigating bias in machinelearning … consequences of machinelearning algorithms and …
… bias concerns unlawful bias, that is, the targeting of people based on prohibited grounds. This may be a subset of ethical bias, but sometimes bias … the basis of gender may be prohibited …
… This section aims to measure genderbias on different multilingual architectures to understand their difference when encoding source sentences and decoding target ones. We will focus …
… may also have been shaded by racial 5 and genderbias,'6 and ultimately reflected in her marks or reference letters from educators.Similar genderbiases may have influenced her male …
A Field, Y Tsvetkov - arXiv preprint arXiv:2004.08361, 2020 - arxiv.org
… While our work serves as an initial approach toward unsupervised detection of comment-level genderbias, we identify several limitations and areas for future work. We first focus on …
… a genderbiased corpus, structured models such as conditional random fields, amplify the bias. … co-related in vSRL, we define the genderbias toward man for each verb b(verb,man) as: …
T Baer, V Kamalnath - McKinsey Insights, 2017 - mckinsey.de
… of bias In automated business processes, machinelearning … Machinelearning also promises to improve decision quality, … biases of past human decision makers, such as gender …