Learning Extremal Representations with Deep Archetypal Analysis

  title={Learning Extremal Representations with Deep Archetypal Analysis},
  author={Sebastian Mathias Keller and Maxim Samarin and Fabricio Arend Torres and Mario Wieser and Volker Roth},
  journal={International Journal of Computer Vision},
  pages={805 - 820}
Archetypes represent extreme manifestations of a population with respect to specific characteristic traits or features. In linear feature space, archetypes approximate the data convex hull allowing all data points to be expressed as convex mixtures of archetypes. As mixing of archetypes is performed directly on the input data, linear Archetypal Analysis requires additivity of the input, which is a strong assumption unlikely to hold e.g. in case of image data. To address this problem, we propose… 

Consistency of Archetypal Analysis

This paper proves a consistency result that shows if the data is independently sampled from a probability measure with bounded support, then the archetype points converge to a solution of the continuum version of the problem, of which it identifies and establishes several properties.

Learning Invariances with Generalised Input-Convex Neural Networks

A novel and exible class of neural networks that generalise input-convex networks that represent functions that are guaranteed to have connected level sets forming smooth manifolds on the input space are introduced.

Learning Conditional Invariance through Cycle Consistency

This work proposes a novel approach to cycle consistency based on the deep information bottleneck and, in contrast to other approaches, allows using continuous target properties and provides inherent model selection capabilities.

Inverse Learning of Symmetries

This work proposes to learn the symmetry transformation with a model consisting of two latent subspaces, where the first subspace captures the target and the second subspace the remaining invariant information, based on the deep information bottleneck in combination with a continuous mutual information regulariser.

3DMolNet: A Generative Network for Molecular Structures

This work proposes a new approach to efficiently generate molecular structures that are not restricted to a fixed size or composition, based on the variational autoencoder which learns a translation-, rotation-, and permutation-invariant low-dimensional representation of molecules.

Neural ADMIXTURE: rapid population clustering with autoencoders

Neural ADMIXTURE is presented, a neural network autoencoder that follows the same modeling assumptions as AD MIXTURE, providing similar (or better) clustering, while reducing the compute time by orders of magnitude.

Inverse Learning of Symmetry Transformations

This work proposes learning two latent subspaces, where the first subspace captures the property and the second subspace the remaining invariant information, based on the deep information bottleneck principle in combination with a mutual information regulariser.

Non-linear archetypal analysis of single-cell RNA-seq data by deep autoencoders

The ability of scAAnet to extract biologically meaningful GEPs using publicly available scRNA-seq datasets including a pancreatic islet dataset, a lung idiopathic pulmonary fibrosis dataset and a prefrontal cortex dataset is demonstrated.

Self-Supervised Representation Learning for High-Content Screening

A self-supervised triplet network is used to learn a phenotypic embedding which is used for visual inspection and top-down assay quality control and outperforms state-of-the-art unsupervised and supervised approaches.

Archetypal Analysis of Geophysical Data illustrated by Sea Surface Temperature

The ability to find and recognize patterns in high-dimensional geophysical data is fundamental to climate science and critical for meaningful interpretation of weather and climate processes.



Archetypal analysis for machine learning and data mining

Coresets for Archetypal Analysis

Empirically, it is shown that efficient coresets for archetypal analysis lead to improved performance on several data sets and theoretical guarantees are derived by showing that quantization errors of k-means upper bound archetyPal analysis can be performed in only two passes over the data.

Automatic Model Selection in Archetype Analysis

This paper presents a novel extension to existing methods for archetype analysis with the specific focus of relaxing the need to provide a fixed number of archetypes beforehand, and introduces a Group-Lasso component popular for sparse linear regression.

Probabilistic archetypal analysis

This paper revisits archetypal analysis from the basic principles, and proposes a probabilistic framework that accommodates other observation types such as integers, binary, and probability vectors that corroborate the proposed methodology with convincing real-world applications.

Unsupervised Learning of Artistic Styles with Archetypal Style Analysis

An unsupervised learning approach to automatically dis- cover, summarize, and manipulate artistic styles from large collections of paintings based on archetypal analysis, which is an un supervised learning technique akin to sparse coding with a geometric interpretation.

Copula Archetypal Analysis

An extension of classical archetypal analysis that is motivated by the observation that classical AA is not invariant against strictly monotone increasing transformations is presented, and introduces a semi-parametric Gaussian copula.

Introduction to archetypal analysis of spatio-temporal dynamics

Archetypal analysis of galaxy spectra

ABSTRACT Archetypal analysis represents each individual member of a set of data vectors as a mixture (aconstrained linear combination)of the pure types or archetypes of the data set. The

Learning Sparse Latent Representations with the Deep Copula Information Bottleneck

This paper adopts the deep information bottleneck model, identifies its shortcomings and proposes a model that circumvents them, and applies a copula transformation which restores the invariance properties of the information bottleneck method and leads to disentanglement of the features in the latent space.

Making Archetypal Analysis Practical

This paper introduces ways of notably accelerating archetypal analysis and introduces the first successful application of the technique to large scale data analysis problems.