LMR Toolbox

This repository serves as the central documentation and tooling hub for the LEMUR Doctoral Network on Learning with Multiple Representations.

It contains:

• Instructions and links to the individual projects' repositories
• Scripts to clone and run experiments locally
• Documentation about the project goals and methodology

---

Project Overview

LEMUR is an MSCA Doctoral Network focused on developing the theoretical foundations, algorithms, and applications of Learning with Multiple Representations (LMR). The core goal is to train 10 PhD researchers in novel approaches to machine learning that emphasize explainability, fairness, weak supervision, and environmental and social impact considerations.

---

Project Objectives

LEMUR’s goals are divided into three major facets:

1. Foundations (F1) — Develop formal guarantees and performance predictions for LMR.
2. Methods (F2) — Create novel LMR algorithms, metrics, and ethical impact evaluations.
3. Applications (F3) — Apply LMR techniques to real-world domains via collaborations with industry partners.

---

Structure

• projects/ — Per-project summary with links to the respective repositories
• scripts/ — Helper scripts to clone and run project repos locally
• config/ — Config files consumed by scripts

---

Quickstart

1. Clone this repo:

   git clone https://github.com/Lemur-dn/lmr-toolbox
   cd lmr-toolbox
   

2. Review participants/ to see links to individual project repositories.

3. Run:

   ./scripts/clone_all.sh
   

---

Contributing to the LMR Toolbox Repo

This repository accepts contributions in:

• Adding new helper scripts in scripts/
• Updating repo links and project info in projects/

Please follow the general GitHub workflow:

1. Fork the repo
2. Create a feature branch
3. Submit a pull request

Be respectful and inclusive — this is a collaborative international research project.

---

Acknowledgements

This project is funded under the Marie Skłodowska-Curie Actions (Horizon Europe).