Character-based recurrent neural networks for morphological relational reasoning
Paper in proceeding, 2017

We present a model for predicting word forms based on morphological relational reasoning with analogies. While previous work has explored tasks such as morphological inflection and reinflection, these models rely on an explicit enumeration of morphological features, which may not be available in all cases. To address the task of predicting a word form given a demo relation (a pair of word forms) and a query word, we devise a character-based recurrent neural network architecture using three separate encoders and a decoder. We also investigate a multiclass learning setup, where the prediction of the relation type label is used as an auxiliary task. Our results show that the exact form can be predicted for English with an accuracy of 94.7%. For Swedish, which has a more complex morphology with more inflectional patterns for nouns and verbs, the accuracy is 89.3%. We also show that using the auxiliary task of learning the relation type speeds up convergence and improves the prediction accuracy for the word generation task.

Author

Olof Mogren

Chalmers, Computer Science and Engineering (Chalmers), Computing Science (Chalmers)

Richard Johansson

University of Gothenburg

EMNLP 2017 - 1st Workshop on Subword and Character Level Models in NLP, SCLeM 2017 - Proceedings of the Workshop

57-63
9781945626913 (ISBN)

EMNLP 2017 1st Workshop on Subword and Character Level Models in NLP, SCLeM 2017
Copenhagen, Denmark,

Subject Categories

Language Technology (Computational Linguistics)

General Language Studies and Linguistics

Bioinformatics (Computational Biology)

More information

Latest update

12/30/2021