Understanding the Political Ideology of Legislators from Social Media Images

Xi, Nan Miles; Ma, Di; Liou, Marcus; Steinert-Threlkeld, Zachary C.; Anastasopoulos, Jason; Joo, Jungseock

doi:10.48550/arxiv.1907.09594

Cited by 5 publications

(2 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We measured counterfactual fairness of commercial computer vision APIs which provide label classification for a large number of visual concepts, including Google Vision API, Amazon Rekognition, IBM Watson Visual Recognition, and Clarifai. These APIs are widely used in commercial products as well as academic research (Xi et al, 2019). While public computer vision datasets usually focus on general concepts (e.g., 60 common object categories in MS COCO (Lin et al, 2014)), these services generate very specific and detailed labels on thousands of distinct concepts.…”

Section: Experiments Computer Vision Apismentioning

confidence: 99%

Gender Slopes: Counterfactual Fairness for Computer Vision Models by Attribute Manipulation

Joo,

Kärkkäinen

2020

Preprint

Self Cite

View full text Add to dashboard Cite

Automated computer vision systems have been applied in many domains including security, law enforcement, and personal devices, but recent reports suggest that these systems may produce biased results, discriminating against people in certain demographic groups. Diagnosing and understanding the underlying true causes of model biases, however, are challenging tasks because modern computer vision systems rely on complex black-box models whose behaviors are hard to decode. We propose to use an encoder-decoder network developed for image attribute manipulation to synthesize facial images varying in the dimensions of gender and race while keeping other signals intact. We use these synthesized images to measure counterfactual fairness of commercial computer vision classifiers by examining the degree to which these classifiers are affected by gender and racial cues controlled in the images, e.g., feminine faces may elicit higher scores for the concept of nurse and lower scores for STEM-related concepts. We also report the skewed gender representations in an online search service on profession-related keywords, which may explain the origin of the biases encoded in the models.

show abstract

Section: Experiments Computer Vision Apismentioning

confidence: 99%

Gender Slopes: Counterfactual Fairness for Computer Vision Models by Attribute Manipulation

Joo,

Kärkkäinen

2020

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Using off-the-shelf tools [2,4] and commercial services, social scientists, who traditionally didn't use images, begun to use images of people to infer their demographic attributes and analyze their behaviors in many studies. Notable examples are demographic analyses of social media users using their photographs [9,43,64,65,62]. The cost of unfair classification is huge as it can over-or under-estimate specific sub-populations in their analysis, which may have policy implications.…”

Section: Face Attribute Recognitionmentioning

confidence: 99%

FairFace: Face Attribute Dataset for Balanced Race, Gender, and Age

Kärkkäinen,

Joo

2019

Preprint

Self Cite

View full text Add to dashboard Cite

Existing public face datasets are strongly biased toward Caucasian faces, and other races (e.g., Latino) are significantly underrepresented. This can lead to inconsistent model accuracy, limit the applicability of face analytic systems to non-White race groups, and adversely affect research findings based on such skewed data. To mitigate the race bias in these datasets, we construct a novel face image dataset, containing 108,501 images, with an emphasis of balanced race composition in the dataset. We define 7 race groups: White, Black, Indian, East Asian, Southeast Asian, Middle East, and Latino. Images were collected from the YFCC-100M Flickr dataset and labeled with race, gender, and age groups. Evaluations were performed on existing face attribute datasets as well as novel image datasets to measure generalization performance. We find that the model trained from our dataset is substantially more accurate on novel datasets and the accuracy is consistent between race and gender groups. The dataset will be released via {https://github.com/joojs/fairface}.

show abstract

Seeing “Us” and “Them”: How Political Symbols Polarize Through Anger, Anxiety, and Enthusiasm

Overgaard,

Coleman

2024

Mass Communication and Society

View full text Add to dashboard Cite

Understanding the Political Ideology of Legislators from Social Media Images

Cited by 5 publications

References 0 publications

Gender Slopes: Counterfactual Fairness for Computer Vision Models by Attribute Manipulation

Gender Slopes: Counterfactual Fairness for Computer Vision Models by Attribute Manipulation

FairFace: Face Attribute Dataset for Balanced Race, Gender, and Age

Seeing “Us” and “Them”: How Political Symbols Polarize Through Anger, Anxiety, and Enthusiasm

Contact Info

Product

Resources

About