discretize_distributions: Efficient Quantization of Gaussian Mixtures with Guarantees in Wasserstein Distance
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866909914138411008 |
|---|---|
| author | Adams, Steven Alwash, Elize Laurenti, Luca |
| author_facet | Adams, Steven Alwash, Elize Laurenti, Luca |
| contents | We present discretize_distributions, a Python package that efficiently constructs discrete approximations of Gaussian mixture distributions and provides guarantees on the approximation error in Wasserstein distance. The package implements state-of-the-art quantization methods for Gaussian mixture models and extends them to improve scalability. It further integrates complementary quantization strategies such as sigma-point methods and provides a modular interface that supports custom schemes and integration into control and verification pipelines for cyber-physical systems. We benchmark the package on various examples, including high-dimensional, large, and degenerate Gaussian mixtures, and demonstrate that discretize_distributions produces accurate approximations at low computational cost. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2511_15854 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | discretize_distributions: Efficient Quantization of Gaussian Mixtures with Guarantees in Wasserstein Distance Adams, Steven Alwash, Elize Laurenti, Luca Machine Learning We present discretize_distributions, a Python package that efficiently constructs discrete approximations of Gaussian mixture distributions and provides guarantees on the approximation error in Wasserstein distance. The package implements state-of-the-art quantization methods for Gaussian mixture models and extends them to improve scalability. It further integrates complementary quantization strategies such as sigma-point methods and provides a modular interface that supports custom schemes and integration into control and verification pipelines for cyber-physical systems. We benchmark the package on various examples, including high-dimensional, large, and degenerate Gaussian mixtures, and demonstrate that discretize_distributions produces accurate approximations at low computational cost. |
| title | discretize_distributions: Efficient Quantization of Gaussian Mixtures with Guarantees in Wasserstein Distance |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2511.15854 |