An unsupervised approach for text summarization for legal documents in English and French.

Approach - Uses clustering of sentence embeddings (generated using Skip-thought model and BERT multilingual model) to generate summarization.