The Dark Side of Dataset Scaling: Evaluating Racial Classification in Multimodal Models.
Abeba BirhaneSepehr DehdashtianVinay Uday PrabhuVishnu BoddetiPublished in: CoRR (2024)
Keyphrases
- model selection
- classification models
- benchmark datasets
- multi modal
- classification accuracy
- outdoor images
- feature vectors
- feature extraction
- decision trees
- decision rules
- pattern recognition
- classification systems
- classification scheme
- machine learning methods
- classification method
- complex systems
- machine learning
- training samples
- text classification
- feature set
- training set
- classification process
- fold cross validation
- training dataset
- database
- neural network
- statistical models
- support vector machine
- probabilistic model
- prior knowledge
- feature space