Probing Classifiers are Unreliable for Concept Removal and Detection.
Abhinav KumarChenhao TanAmit SharmaPublished in: CoRR (2022)
Keyphrases
- automatic detection
- detection algorithm
- support vector
- false alarms
- support vector machine classifier
- classification systems
- concept detection
- detection accuracy
- concept learning
- classification models
- detection method
- naive bayes
- discriminative classifiers
- object detectors
- machine learning methods
- detection rate
- object detection
- training set
- training data
- decision trees
- genetic algorithm
- boosted classifiers
- linear classifiers
- ensemble learning
- classification algorithm
- svm classifier
- machine learning algorithms
- feature selection
- computer vision