Saved in:
Bibliographic Details
Main Authors: Tesfazgi, Samuel, Sprandl, Leonhard, Hirche, Sandra
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2504.02607
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909563440070656
author Tesfazgi, Samuel
Sprandl, Leonhard
Hirche, Sandra
author_facet Tesfazgi, Samuel
Sprandl, Leonhard
Hirche, Sandra
contents The practical deployment of learning-based autonomous systems would greatly benefit from tools that flexibly obtain safety guarantees in the form of certificate functions from data. While the geometrical properties of such certificate functions are well understood, synthesizing them using machine learning techniques still remains a challenge. To mitigate this issue, we propose a diffeomorphic function learning framework where prior structural knowledge of the desired output is encoded in the geometry of a simple surrogate function, which is subsequently augmented through an expressive, topology-preserving state-space transformation. Thereby, we achieve an indirect function approximation framework that is guaranteed to remain in the desired hypothesis space. To this end, we introduce a novel approach to construct diffeomorphic maps based on RBF networks, which facilitate precise, local transformations around data. Finally, we demonstrate our approach by learning diffeomorphic Lyapunov functions from real-world data and apply our method to different attractor systems.
format Preprint
id arxiv_https___arxiv_org_abs_2504_02607
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Learning Geometrically-Informed Lyapunov Functions with Deep Diffeomorphic RBF Networks
Tesfazgi, Samuel
Sprandl, Leonhard
Hirche, Sandra
Machine Learning
Artificial Intelligence
Systems and Control
The practical deployment of learning-based autonomous systems would greatly benefit from tools that flexibly obtain safety guarantees in the form of certificate functions from data. While the geometrical properties of such certificate functions are well understood, synthesizing them using machine learning techniques still remains a challenge. To mitigate this issue, we propose a diffeomorphic function learning framework where prior structural knowledge of the desired output is encoded in the geometry of a simple surrogate function, which is subsequently augmented through an expressive, topology-preserving state-space transformation. Thereby, we achieve an indirect function approximation framework that is guaranteed to remain in the desired hypothesis space. To this end, we introduce a novel approach to construct diffeomorphic maps based on RBF networks, which facilitate precise, local transformations around data. Finally, we demonstrate our approach by learning diffeomorphic Lyapunov functions from real-world data and apply our method to different attractor systems.
title Learning Geometrically-Informed Lyapunov Functions with Deep Diffeomorphic RBF Networks
topic Machine Learning
Artificial Intelligence
Systems and Control
url https://arxiv.org/abs/2504.02607