Home › Categories › Technology› Bias In AI Systems

Bias In AI Systems.pdf

1908.09635.pdf

🔗 Source: arxiv.org

📊 Size: 1.63 MB

👤 Author: Ninareh Mehrabi, Fred Morstatter, Nripsuta Saxena, Kristina Lerman, and Aram Galstyan, USC-ISI

⬇️ Downloads: 363

Summary

Machine learning algorithms are vulnerable to biases, rendering their decisions unfair, with examples including a tool used by US courts to make pretrial detention decisions that was biased against African-Americans, and an AI system that judged beauty pageant winners but was biased against darker-skinned contestants. Biases can arise from data and algorithms, and unfair algorithmic outcomes can impact user experience, generating a feedback loop that perpetuates existing sources of bias. Researchers have identified various sources of biases, including biases in data and algorithms, and have developed different fairness definitions and approaches to address these issues in different machine learning domains.

Download File

Description

Machine learning algorithms are vulnerable to biases, rendering their decisions unfair, with examples including a tool used by US courts to make pretrial...

Technical Information

File Format: PDF
File Size: 1.63 MB
Pages: 34
Language: EN
Author: Ninareh Mehrabi, Fred Morstatter, Nripsuta Saxena, Kristina Lerman, and Aram Galstyan, USC-ISI
Total Downloads: 363
Last Updated: 4 weeks ago

Document Overview

This PDF document about Bias in AI Systems provides comprehensive information and guidance. Whether you're a beginner or advanced user, this resource offers valuable insights into Bias in AI Systems.