Unsupervised learning of phase transitions: from principal component analysis to variational autoencoders.
Sebastian Johann WetzelPublished in: CoRR (2017)
Keyphrases
- phase transition
- unsupervised learning
- principal component analysis
- dimensionality reduction
- deep belief networks
- constraint satisfaction
- denoising
- combinatorial problems
- supervised learning
- random constraint satisfaction problems
- principal components
- satisfiability problem
- independent component analysis
- np complete
- image segmentation
- semi supervised
- random instances
- object recognition
- randomly generated
- face recognition
- graph coloring
- cellular automata
- deep learning
- sat problem
- hard problems
- feature space
- feature extraction
- expectation maximization
- np complete problems
- text classification
- restricted boltzmann machine
- image processing
- feature selection
- training data
- objective function
- constraint programming
- constraint satisfaction problems
- machine learning
- training set
- high dimensional
- data points