Key points are not available for this paper at this time.
Cross-lingual annotations of legislative texts enable us to explore major covered in multilingual legal data and are a key facilitator of semantic when searching for similar documents. Multilingual probabilistic models have recently emerged as a group of semi-supervised machine models that can be used to perform thematic explorations on of texts in multiple languages. However, these approaches require-aligned training data to create a language-independent space, which the amount of scenarios where this technique can be used. In this work, provide an unsupervised document similarity algorithm based on hierarchies multi-lingual concepts to describe topics across languages. The algorithm not require parallel or comparable corpora, or any other type of resource. Experiments performed on the English, Spanish, French and editions of JCR-Acquis corpora reveal promising results on and sorting documents by similar content.
Badenes-Olmedo et al. (Thu,) studied this question.