R
Richa Singh
Researcher at Indian Institute of Technology, Jodhpur
Publications - 493
Citations - 11353
Richa Singh is an academic researcher from Indian Institute of Technology, Jodhpur. The author has contributed to research in topics: Facial recognition system & Deep learning. The author has an hindex of 53, co-authored 422 publications receiving 9145 citations. Previous affiliations of Richa Singh include Indian Institutes of Technology & University of Virginia.
Papers
More filters
Proceedings ArticleDOI
Misclassifications of Contact Lens Iris PAD Algorithms: Is it Gender Bias or Environmental Conditions?
TL;DR: In this article , the authors presented a rigorous study on gender bias in iris presentation attack detection algorithms using a large-scale and gender-balanced database, which can help in building future presentation attacks detection algorithms with the aim of fair treatment of each demography.
Journal ArticleDOI
Uniform Misclassification Loss for Unbiased Model Prediction
TL;DR: In this article , the authors proposed a novel loss function, termed as Uniform Misclassification Loss (UML), to train deep models for unbiased outcomes, which penalizes the model for the worst performing subgroup for mitigating bias and enhancing the overall model performance.
Journal ArticleDOI
Socio-economic Status of Female Workers Engaged in Traditional Chikankari under Sitapur District
Richa Singh,Purnima Shah +1 more
TL;DR: In this paper , the Modified Kuppuswamy Socio-economic scale was used to calculate the socio-economic status of the respondents in Chikankari work and found that the workers were mostly upper lower class (60%) and lower middle class (40%).
Journal ArticleDOI
Detox Loss: Fairness Constraints for Learning With Imbalanced Data
TL;DR: In this article , the authors proposed Detox loss, a bias invariant feature learning loss function for learning unbiased models, which can be used to learn fairer deep learning classifiers, and mitigate bias from existing pre-trained networks, especially in the challenging constraint of imbalanced training data with respect to a protected attribute.