Elaine Zosa author Mark Granroth-Wilding author Lidia Pivovarova author 2020-may text English eng European Language Resources Association Marseille, France conference publication 979-10-95546-55-9 We address the problem of linking related documents across languages in a multilingual collection. We evaluate three diverse unsupervised methods to represent and compare documents: (1) multilingual topic model; (2) cross-lingual document embeddings; and (3) Wasserstein distance.We test the performance of these methods in retrieving news articles in Swedish that are known to be related to a given Finnish article.The results show that ensembles of the methods outperform the stand-alone methods, suggesting that they capture complementary characteristics of the documents zosa-etal-2020-comparison https://www.aclweb.org/anthology/2020.clssts-1.6 2020-may 32 37