https://github.com/ills-montreal/emir

When is an Embedding Model More Promising than Another?, NeurIPS'24
https://github.com/ills-montreal/emir

embedders information-theory molecule nlp representation-learning

Last synced: about 2 months ago
JSON representation

When is an Embedding Model More Promising than Another?, NeurIPS'24

• Host: GitHub
• URL: https://github.com/ills-montreal/emir
• Owner: ills-montreal
• Created: 2023-10-27T18:43:58.000Z (about 1 year ago)
• Default Branch: main
• Last Pushed: 2024-11-05T20:09:56.000Z (about 2 months ago)
• Last Synced: 2024-11-05T20:35:04.597Z (about 2 months ago)
• Topics: embedders, information-theory, molecule, nlp, representation-learning
• Language: Jupyter Notebook
• Homepage: https://arxiv.org/pdf/2406.07640v1
• Size: 296 MB
• Stars: 6
• Watchers: 3
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: Readme.md

Awesome Lists containing this project

README

        
# When is an Embedding Model More Promising than Another?

**NeurIPS'24**

*Maxime DARRIN\*, Philippe FORMONT\*, Ismail BEN AYED, Jackie Chi Kit CHEUNG, Pablo PIANTANIDA*

[[ArXiv](https://arxiv.org/abs/2406.07640)]

![image](https://github.com/user-attachments/assets/97125cc0-aa8e-48ee-88d5-7a7201507516)

This repository contains the code for the paper: **When is an Embedding Model More Promising than Another?**

It is divided in three main components:

1. [emir](emir): Contains the code to estimate the information sufficiency between two models.

2. [molecule](molecule): Contains the code used for the molecular modeling experiments.

3. [nlp_embeddings](nlp_embeddings): Contains the code used for the NLP experiments

All directories contain there specific Readme files.

## Introduction

### Installation

```bash

pip install -e . 

```

Cite us

-

```

Coming soon

```