https://github.com/cypriengille/semi-supervised-autoencoder
A sparsified AutoEncoder to solve Semi-Supervised classification tasks
https://github.com/cypriengille/semi-supervised-autoencoder
autoencoder biomedical semisupervised-learning
Last synced: 7 months ago
JSON representation
A sparsified AutoEncoder to solve Semi-Supervised classification tasks
- Host: GitHub
- URL: https://github.com/cypriengille/semi-supervised-autoencoder
- Owner: CyprienGille
- Created: 2022-08-19T13:00:21.000Z (about 3 years ago)
- Default Branch: main
- Last Pushed: 2024-03-12T09:46:45.000Z (over 1 year ago)
- Last Synced: 2025-01-27T06:13:32.288Z (8 months ago)
- Topics: autoencoder, biomedical, semisupervised-learning
- Language: Python
- Homepage:
- Size: 34.1 MB
- Stars: 2
- Watchers: 3
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Semisupervised Autoencoder
This repository contains the code from :
> A new Semi-supervised classification method using a supervised autoencoder for biomedical applications, Gille C. and Guyard F. and Barlaud M., (2022). https://arxiv.org/abs/2208.10315 .
## Repository contents
- `script_semisupervised.py` : This is the main script used to produce the results shown in the paper. It generates plots in the `plots` directory, and saves results (metrics, losses...) as CSVs in the `results_semi` folder. All parameters are tunable near the start of the script.
- `param_plots` : This is a helper script to reproduce the plots from Figures 2 and 3 of the aforementioned paper.
- `script_eta_optimization.py` : This script is used to find the optimal sparsification parameter $\eta$ either by dichotomy or using the [golden section strategy](https://en.wikipedia.org/wiki/Golden-section_search).
- `functions` : Contains function utilities useful for the other main scripts.
- `data` : Contains the two datasets presented in the paper.
- `plots` and `results_semi` are results directories filled by executing `semisupervised_tests.py`.