Skip to content

What is Parallel Sentence Mining?

Parallel sentence mining is a process of searching parallel (translated) sentence pairs in monolingual corpora.

source (English)
1
2
3
This wine bar - restaurant promises you beautiful culinary surprises.
Every cup of coffee should create a personal moment of pleasure.
Some text that is not translated.
target (French)
1
2
3
Chaque tasse de café devrait créer un moment de plaisir personnel.
Deux
Ce bar à vins - restaurant vous promet de belles surprises culinaires.

The goal is to identify all translation pairs between the source and target sets of sentences.

source target index
This wine bar - restaurant promises you beautiful culinary surprises. Ce bar à vins - restaurant vous promet de belles surprises culinaires. 1 - 3
Every cup of coffee should create a personal moment of pleasure. Chaque tasse de café devrait créer un moment de plaisir personnel. 2 - 1