A generative model that learns Betti numbers from a data set
نویسندگان
چکیده
Analysis of multidimensional data is challenging. Topological invariants can be used to summarize essential features of such data sets. In this work, we propose to compute the Betti numbers from a generative model based on a simplicial complex learnt from the data. We compare it to the Witness Complex, a geometric technique based on nearest neighbors. Our results on different data distributions with known topology show that Betti numbers are well recovered with our method.
منابع مشابه
The Generative Simplicial Complex to extract Betti numbers from unlabelled data
Analysis of multidimensional data is challenging. Topological invariants can be used to summarize essential features of such data sets. In this work, we propose to compute the Betti numbers from a generative model based on a simplicial complex learnt from the data. The algorithm called a Generative Simplicial Complex can extract topological features from multidimensional datasets with non-trivi...
متن کاملFast methods for computing isosurface topology with Betti numbers
Betti numbers can be used as a means for feature detection to aid in the exploration of complex large-scale data sets. We present a fast algorithm for the calculation of Betti numbers for triangulated isosurfaces, along with examples of their use. Once an isosurface is extracted from a data set, calculating Betti numbers only requires time and space proportional to the isosurfaces, not the data...
متن کاملOn a special class of Stanley-Reisner ideals
For an $n$-gon with vertices at points $1,2,cdots,n$, the Betti numbers of its suspension, the simplicial complex that involves two more vertices $n+1$ and $n+2$, is known. In this paper, with a constructive and simple proof, wegeneralize this result to find the minimal free resolution and Betti numbers of the $S$-module $S/I$ where $S=K[x_{1},cdots, x_{n}]$ and $I$ is the associated ideal to ...
متن کاملGenerative Vector Quantisation
{ Based on the assumption that a pattern is constructed out of features which are either fully present or absent, we propose a vector quantisation method which constructs patterns using binary combinations of features. For this model there exists an eecient EM-like learning algorithm which learns a set of representative codebook vectors. In terms of a generative model, the collection of allowed...
متن کاملAdversarial Autoencoders
In this paper, we propose the “adversarial autoencoder” (AAE), which is a probabilistic autoencoder that uses the recently proposed generative adversarial networks (GAN) to perform variational inference by matching the aggregated posterior of the hidden code vector of the autoencoder with an arbitrary prior distribution. Matching the aggregated posterior to the prior ensures that generating fro...
متن کامل