Making Time Series Embeddings More Interpretable in Deep Learning - Extracting Higher-Level Features via Symbolic Approximation Representations.
Leonid SchwenkeMartin AtzmuellerPublished in: FLAIRS (2023)
Keyphrases
- deep learning
- unsupervised feature learning
- higher level
- low level
- symbolic representation
- deep belief networks
- higher order
- unsupervised learning
- machine learning
- decision trees
- image features
- high level
- co occurrence
- mental models
- feature vectors
- feature extraction
- input image
- probabilistic model
- restricted boltzmann machine