Abstract

Contextualized word embeddings like BERT enabled significant advances in many natural language processing tasks. Recently, multilingual versions of such embeddings were trained on large text corpora of more than 100 languages. In this paper we investigate how well such embeddings perform in zero-shot cross lingual transfer for an event extraction task. In particular, we analyze the impact of the alignment of contextualized word embeddings using a parallel corpus on the performance of the downstream task.

Date

2020

Authors

Karen Hambardzumyan, Hrant Khachatrian, Jonathan May

Journal

Collaborative Technologies and Data Science in Artificial Intelligence Applications