![]() ![]() Anthology ID: D19-1091 Volume: Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP) Month: November Year: 2019 Address: Hong Kong, China Venues: EMNLP Also, we identify the crucial factors for success with cross-lingual transfer for morphological inflection: typological similarity and a common representation across languages. The macro-averaged accuracy of our models outperforms the state-of-the-art by 15 percentage points. In addition, we investigate the effects of cross-lingual transfer from single and multiple languages, as well as monolingual data hallucination. First, we present a novel two-step attention architecture for the inflection decoder. In response, we propose a battery of improvements that greatly improve performance under such low-resource conditions. ![]() However, for a long tail of languages the necessary resources are hard to come by, and state-of-the-art neural methods that work well under higher resource settings perform poorly in the face of a paucity of data. ![]() Abstract Recent years have seen exceptional strides in the task of automatic morphological inflection generation. ![]()
0 Comments
Leave a Reply. |