https://github.com/mariamagro/exploring_variational_autoencoders_gaussian_mixture_models

This project investigates Variational Autoencoders (VAEs) for generating synthetic data from a 3D Gaussian Mixture Model (GMM). It covers data synthesis, VAE architecture with dense layers, and model evaluation, including clustering and T-SNE visualization to assess the VAE's performance and latent space representation.
https://github.com/mariamagro/exploring_variational_autoencoders_gaussian_mixture_models

gmm vae-pytorch

Last synced: about 1 year ago
JSON representation

This project investigates Variational Autoencoders (VAEs) for generating synthetic data from a 3D Gaussian Mixture Model (GMM). It covers data synthesis, VAE architecture with dense layers, and model evaluation, including clustering and T-SNE visualization to assess the VAE's performance and latent space representation.

• Host: GitHub
• URL: https://github.com/mariamagro/exploring_variational_autoencoders_gaussian_mixture_models
• Owner: mariamagro
• Created: 2024-08-18T21:18:38.000Z (almost 2 years ago)
• Default Branch: main
• Last Pushed: 2024-08-18T21:29:55.000Z (almost 2 years ago)
• Last Synced: 2025-01-30T09:23:17.183Z (over 1 year ago)
• Topics: gmm, vae-pytorch
• Language: Jupyter Notebook
• Homepage:
• Size: 4.5 MB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
# Neural Networks: Exploring Variational Autoencoders with Gaussian Mixture Models

**Credit:**

The implementation is based on tasks provided by Pablo M. Olmos.

**Completed by:**

- **María Ángeles Magro Garrote** - [mariamagro](https://github.com/mariamagro)

- **Marina Gómez Rey** - [MarinaGRey](https://github.com/MarinaGRey)

- **Ángela Durán**

## Overview

This project explores the application of Variational Autoencoders (VAEs) in the context of generating synthetic data from a 3-dimensional Gaussian Mixture Model (GMM). Our aim is to evaluate the VAE's capability to capture the intricate structure of multi-modal distributions and produce samples that closely emulate the ground truth distribution.

The project consists of the following components:

- **Synthetic Data Generation**: We use a Gaussian Mixture Model (GMM) to create a synthetic dataset for training and evaluating the VAE.

- **VAE Architecture**: We construct and train a VAE using dense layers, and analyze its performance in capturing and reconstructing the data.

- **Clustering and Mode Identification**: We use clustering techniques and mode identification to analyze the VAE's latent space and compare it with the ground truth data.

- **T-SNE Visualization**: We visualize the latent space representations using T-SNE to gain insights into the VAE's performance and the structure of the data.

## Contents

- `report.pdf`: A comprehensive report detailing the project, including methodology, results, and insights.

- `part1.ipynb`: Jupyter Notebook containing the implementation of synthetic data generation and VAE model definition.

- `part2.ipynb`: Jupyter Notebook for training the VAE, generating samples, performing clustering, mode identification, and T-SNE visualization.