Published May 3, 2020
| Version v1
Conference paper
Open
A Comparison of Unsupervised Methods for Ad hoc Cross-Lingual Document Retrieval
- 1. University of Helsinki, Helsinki, Finland
Description
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.
Files
Zosa_LREC2020.pdf
Files
(173.6 kB)
Name | Size | Download all |
---|---|---|
md5:70cf052695bb5dd148bb1524994a9e73
|
173.6 kB | Preview Download |