Gender Stereotyping Impact in Facial Expression Recognition

Dominguez-Catena, Iris; Paternain, Daniel; Galar, Mikel

doi:10.1007/978-3-031-23618-1_1

Cited by 6 publications

(14 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This value is independent for different number of represented groups, so that datasets with different R can achieve the same SEI. Normalized Standard Deviation (NSD) [8]. Another metric for evenness is the Normalized Standard Deviation of the population distribution.…”

Section: Metrics For Representational Biasmentioning

confidence: 99%

“…Thus, the application of this metric could potentially establish a differentiation between forward and backward stereotypical bias. Normalized Mutual Information (NMI) [8], [64]. A different approach to measuring the association between two variables is the use of Mutual Information based variables.…”

Section: Metrics For Stereotypical Biasmentioning

confidence: 99%

“…Finally, the stereotypical bias can also be measured using the following local metrics. Normalized Pointwise Mutual Information (NPMI) [8], [64]. The NMI(X) metric has a local variant, NPMI.…”

Section: Metrics For Stereotypical Biasmentioning

confidence: 99%

“…To overcome this limitation, we propose, following our previous work [8], [86], the study of biases with respect to a proxy demographic prediction instead of the original unrecoverable information. This can be achieved through a secondary demographic model, such as the FairFace [84] face model, based on the homonymous dataset.…”

Section: Fairfacementioning

confidence: 99%

“…Although different metrics have been proposed [6], most of them are only focused on the bias exhibited by the trained models. The measurement of bias in the source datasets has not got the same attention, although it enables the validation of new bias mitigation methods [7], the explanation of bias transference throughout the training process [8], and the demographic description of the application environment where a dataset or model can be safely used [9].…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Metrics for Dataset Demographic Bias: A Case Study on Facial Expression Recognition

Dominguez-Catena¹,

Paternain²,

Galar³

2023

Preprint

View full text Add to dashboard Cite

Demographic biases in source datasets have been shown as one of the causes of unfairness and discrimination in the predictions of Machine Learning models. One of the most prominent types of demographic bias are statistical imbalances in the representation of demographic groups in the datasets. In this paper, we study the measurement of these biases by reviewing the existing metrics, including those that can be borrowed from other disciplines. We develop a taxonomy for the classification of these metrics, providing a practical guide for the selection of appropriate metrics. To illustrate the utility of our framework, and to further understand the practical characteristics of the metrics, we conduct a case study of 20 datasets used in Facial Emotion Recognition (FER), analyzing the biases present in them. Our experimental results show that many metrics are redundant and that a reduced subset of metrics may be sufficient to measure the amount of demographic bias. The paper provides valuable insights for researchers in AI and related fields to mitigate dataset bias and improve the fairness and accuracy of AI models. The code is available at https://github.com/irisdominguez/dataset bias metrics.

show abstract

Section: Metrics For Representational Biasmentioning

confidence: 99%

Section: Metrics For Stereotypical Biasmentioning

confidence: 99%

Section: Metrics For Stereotypical Biasmentioning

confidence: 99%

Section: Fairfacementioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Metrics for Dataset Demographic Bias: A Case Study on Facial Expression Recognition

Dominguez-Catena¹,

Paternain²,

Galar³

2023

Preprint

View full text Add to dashboard Cite

show abstract

Less can be more: representational vs. stereotypical gender bias in facial expression recognition

Dominguez-Catena,

Paternain,

Jurio

et al. 2024

Prog Artif Intell

View full text Add to dashboard Cite

Machine learning models can inherit biases from their training data, leading to discriminatory or inaccurate predictions. This is particularly concerning with the increasing use of large, unsupervised datasets for training foundational models. Traditionally, demographic biases within these datasets have not been well-understood, limiting our ability to understand how they propagate to the models themselves. To address this issue, this paper investigates the propagation of demographic biases from datasets into machine learning models. We focus on the gender demographic component, analyzing two types of bias: representational and stereotypical. For our analysis, we consider the domain of facial expression recognition (FER), a field known to exhibit biases in most popular datasets. We use Affectnet, one of the largest FER datasets, as our baseline for carefully designing and generating subsets that incorporate varying strengths of both representational and stereotypical bias. Subsequently, we train several models on these biased subsets, evaluating their performance on a common test set to assess the propagation of bias into the models’ predictions. Our results show that representational bias has a weaker impact than expected. Models exhibit a good generalization ability even in the absence of one gender in the training dataset. Conversely, stereotypical bias has a significantly stronger impact, primarily concentrated on the biased class, although it can also influence predictions for unbiased classes. These results highlight the need for a bias analysis that differentiates between types of bias, which is crucial for the development of effective bias mitigation strategies.

show abstract

Unbiased Decision-Making Framework in Long-Video Macro & Micro-Expression Spotting

Tan,

Rajanala,

Pal

et al. 2023

2023 Asia Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC)

View full text Add to dashboard Cite

Gender Stereotyping Impact in Facial Expression Recognition

Cited by 6 publications

References 26 publications

Metrics for Dataset Demographic Bias: A Case Study on Facial Expression Recognition

Metrics for Dataset Demographic Bias: A Case Study on Facial Expression Recognition

Less can be more: representational vs. stereotypical gender bias in facial expression recognition

Unbiased Decision-Making Framework in Long-Video Macro & Micro-Expression Spotting

Contact Info

Product

Resources

About