Why Overfitting Isn't Always Bad: Retrofitting Cross-Lingual Word Embeddings to Dictionaries

• Published in: arXiv.org
• Main Authors: Zhang, Mozhi, Fujinuma, Yoshinari, Paul, Michael J, Boyd-Graber, Jordan
• Format: Paper
• Language: English
• Published: Ithaca Cornell University Library, arXiv.org 01.05.2020
• Subjects: Dictionaries; Downstream effects; Post-processing; Retrofitting; Training

Cross-lingual word embeddings (CLWE) are often evaluated on bilingual lexicon induction (BLI). Recent CLWE methods use linear projections, which underfit the training dictionary, to generalize on BLI. However, underfitting can hinder generalization to other downstream tasks that rely on words from t...