Characterizing High-dimensional Dynamics by Combinatorial-Topological Methods on a Latent Space
Abstract
Combinatorial-topological methods for characterizing dynamics are rigorous, generalizable, computable, and they only require approximations, but the dimension of the phase space is a computational bottleneck to their wider application. Motivated by the growing number of machine learning techniques for obtaining lower-dimensional latent representations of dynamics, we present an initial study of combinatorial-topological techniques in the dimensionality reduction setting. We establish bounds under which an algebraic structure that organizes dynamics can be lifted from the latent space to the original system. As a corollary, one can conclude the existence of attractors within certain regions of the original phase space. The hypothesis of these results is expressed in terms of an approximate semiconjugacy between the original and latent dynamics. To demonstrate the ideas, we combine autoencoder-based models with Conley-Morse graph computations for Leslie population models, a thirteen-dimensional Mediterranean red coral population model, and the Chafee--Infante equation. While the lift of the Conley index is still an open question, the examples recover the expected algebraic topological invariants in several settings.
Related
- Ramsey theory and topological dynamics of 0-dimensional flows
- Topological Insights and Combinatorial Frameworks for Twist Knot Complements
- A Combinatorial Benders Decomposition Framework for Two-Dimensional Irregular Bin Packing Problems with Convex Polygons
- Characterizing higher-dimensional Hölder domains via spherical geometric functions
- Characterizing LLM scientific concept generation: A multi-dimensional measurement study
BibTeX
@misc{bailon2026characterizing,
title = {{Characterizing High-dimensional Dynamics by Combinatorial-Topological Methods on a Latent Space}},
author = {Patrick Bailon and Marcio Gameiro and Brittany Gelb and William Kalies and Miroslav Kramar and Konstantin Mischaikow and Bernardo Rivas and Ewerton Vieira},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.01509},
eprint = {2609.01509},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.01509},
}
RIS
TY - UNPB
TI - Characterizing High-dimensional Dynamics by Combinatorial-Topological Methods on a Latent Space
AU - Patrick Bailon
AU - Marcio Gameiro
AU - Brittany Gelb
AU - William Kalies
AU - Miroslav Kramar
AU - Konstantin Mischaikow
AU - Bernardo Rivas
AU - Ewerton Vieira
PY - 2026
DA - 2026/09/01
PB - arXiv
DO - 10.48550/arxiv.2609.01509
UR - https://arxiv.org/abs/2609.01509
AB - Combinatorial-topological methods for characterizing dynamics are rigorous, generalizable, computable, and they only require approximations, but the dimension of the phase space is a computational bottleneck to their wider application. Motivated by the growing number of machine learning techniques for obtaining lower-dimensional latent representations of dynamics, we present an initial study of combinatorial-topological techniques in the dimensionality reduction setting. We establish bounds under which an algebraic structure that organizes dynamics can be lifted from the latent space to the original system. As a corollary, one can conclude the existence of attractors within certain regions of the original phase space. The hypothesis of these results is expressed in terms of an approximate semiconjugacy between the original and latent dynamics. To demonstrate the ideas, we combine autoencoder-based models with Conley-Morse graph computations for Leslie population models, a thirteen-dimensional Mediterranean red coral population model, and the Chafee--Infante equation. While the lift of the Conley index is still an open question, the examples recover the expected algebraic topological invariants in several settings.
ER -