Deep Learning on Small Datasets without Pre-Training using Cosine Loss.
Björn BarzJoachim DenzlerPublished in: WACV (2020)
Keyphrases
- deep learning
- restricted boltzmann machine
- deep architectures
- unsupervised learning
- unsupervised feature learning
- deep belief networks
- machine learning
- mental models
- supervised learning
- training set
- weakly supervised
- object recognition
- high dimensional
- viewpoint
- small number
- object detection
- training examples
- multiscale
- bayesian networks
- decision trees
- data sets