Autoencoding topology
arXiv:1803.00156
Abstract
The problem of learning a manifold structure on a dataset is framed in terms of a generative model, to which we use ideas behind autoencoders (namely adversarial/Wasserstein autoencoders) to fit deep neural networks. From a machine learning perspective, the resulting structure, an atlas of a manifold, may be viewed as a combination of dimensionality reduction and "fuzzy" clustering.
arXiv comment10 pages
References in corpus (1)
Cited by in corpus (5)
- Discovering physical concepts with neural networks
- Change Detection in Graph Streams by Learning Graph Embeddings on Constant-Curvature Manifolds
- Adversarial Autoencoders with Constant-Curvature Latent Manifolds
- A Topological Framework for Deep Learning
- Atlas Based Representation and Metric Learning on Manifolds