On the use of binary stochastic autoencoders for multi-label classification under the zero-one loss.
Denis LecoeucheAlex AussemMaxime GassePublished in: INNS Conference on Big Data (2018)
Keyphrases
- multi label classification
- binary classifiers
- multi label
- multi class
- binary classification
- image annotation
- denoising
- semi supervised learning
- text categorization
- label space
- text classification
- multi class classification
- image classification
- classification algorithm
- semi supervised
- similarity measure
- graph cuts
- meta level
- hyperplane
- learning tasks
- learning algorithm
- training data
- pairwise
- probabilistic model
- cross validation
- class labels
- visual features
- multi modal