Feb. 13, 2024, 9:51 p.m. | Kaggle

Kaggle www.youtube.com

About the project:
Twitter once automatically cropped user-uploaded images to fit a specific aspect ratio. However, Twitter users noticed that this automatic cropping (which is saliency-based) tended to favor certain groups of individuals, for example those with lighter skin. My project quantifies this bias and builds upon the paper ‘Auditing saliency cropping algorithms’ by Birthane, Prabhu and Whaley. My notebook: 1) creates thousands of images with ordered pairs of two individuals of different self-identified racial/ethnic/gender backgrounds 2) runs these generated …

bias example images issa kaggle paper presentation project twitter

Founding AI Engineer, Agents

@ Occam AI | New York

AI Engineer Intern, Agents

@ Occam AI | US

AI Research Scientist

@ Vara | Berlin, Germany and Remote

Data Architect

@ University of Texas at Austin | Austin, TX

Data ETL Engineer

@ University of Texas at Austin | Austin, TX

Lead GNSS Data Scientist

@ Lurra Systems | Melbourne