SINr: a python package to train interpretable word and graph embeddings - Archive ouverte HAL Access content directly
Conference Papers Year : 2023

SINr: a python package to train interpretable word and graph embeddings

Abstract

In this paper, we introduce the SINr Python package to train word and graph embeddings. The SINr approach is based on community detection: a vector for a node is built upon the distribution of its connections through the communities detected on the graph at hand. Because of this, the algorithm runs very fast, and does not require GPUs to proceed. Furthermore, the dimensions of the embedding space are interpretable, those are based on the communities extracted. The package is distributed under Cecill-2.1 license and is available on Github and pypi.
Fichier principal
Vignette du fichier
FRCSS_2023___SINr_library-2.pdf (232.24 Ko) Télécharger le fichier
Origin : Files produced by the author(s)
Licence : CC BY - Attribution

Dates and versions

hal-04113024 , version 1 (01-06-2023)

Licence

Attribution

Identifiers

Cite

Thibault Prouteau, Nicolas Dugué, Simon Guillot, Anthony Perez. SINr: a python package to train interpretable word and graph embeddings. French Regional Conference on Complex Systems, May 2023, Le Havre, France. pp.215, ⟨10.5281/zenodo.7957531⟩. ⟨hal-04113024⟩
37 View
30 Download

Altmetric

Share

Gmail Facebook X LinkedIn More