Cross-lingual Semantic Memory

Restict result to language: Show known translations in: Restrict to memory:

XSM is a Cross-lingual Semantic Memory. The current version is fed with 80M sentences, in 4 languages (French, English, German and, Spanish), from various corpus: MultiUN, Europarl, EMEA, DGT, ECB, JRC-Acquis’, OpenSubtitle, Wikipedia, Gnome, KDE4, Ubuntu, OpenOffice, PHP, ... You may query it in any language to find sentences having similar meaning in any language stored in the XSM database. Your query sentence may be in a language not already imported in the XSM database.

The translation part is built using triangular pairings, thus it also provides with source-to-source language pairs:


Clusterisation/indexation sémantique multi-résolution
80M de phrases (FR, EN, DE, ES)
Projection sur les axes 82/83 en dimension 1024

Adaptation d'un corpus générique à un corpus technique
80M de phrases (FR, EN, DE, ES)
Projection sur les axes 94/95 en dimension 1024 Stay informed on Twitter @NeuroSpellO cub[AI]x