LITE: A Framework for Lattice-Integrated Embedding of Topological Descriptors

Author: Michael Etienne Van Huffel

Description

Implementation of Lattice Integrated Topological Embedding (LITE) for Persistence Diagrams based on the paper https://arxiv.org/abs/2312.17093.

Overview

This repository contains the official implementation of LITE, a method for embedding Persistence Diagrams into elements of vetor spaces. It showcases the integration of such vectorization technique in the context of graph classification and dynamical particles classification.

Repository Structure

The repository is systematically organized to facilitate easy navigation and comprehensive understanding of each component.

Jupyter Notebooks

• graph-classy.ipynb: Demo that demonstrates the application of LITE Vectorization in graph classification.
• orbit-classy.ipynb: Demo that demonstrates the application of LITE Vectorization in dynamical particles classification.

Python Scripts

• lite/lite.py: Core implementation of the LITE Vectorization algorithm.
• lite/utils.py: Provides essential utility functions for data processing and analysis within the notebooks.

Installation

To reproduce the analysis environment, you will need Python 3.6 or later. Please install the required Python packages listed in requirements.txt.

git clone git@github.com:majkevh/spectral-master.git
cd spectral-master
pip install -r requirements.txt

Data Folder

• data/: Contains all datasets used in the notebooks. For conducting experiments involving graph data, I utilized datasets and functions sourced from the PersLay and ATOL repositories.

License

This project is licensed under the MIT License - see the LICENSE file for details.