Paper: Robust multilingual statistical morphological generation models

ACL ID P13-3023
Title Robust multilingual statistical morphological generation models
Venue Annual Meeting of the Association of Computational Linguistics
Session Student Session
Year 2013
Authors

We present a novel method of statisti- cal morphological generation, i.e. the pre- diction of inflected word forms given lemma, part-of-speech and morphological features, aimed at robustness to unseen in- puts. Our system uses a trainable classifier to predict ?edit scripts? that are then used to transform lemmas into inflected word forms. Suffixes of lemmas are included as features to achieve robustness. We evalu- ate our system on 6 languages with a vary- ing degree of morphological richness. The results show that the system is able to learn most morphological phenomena and gen- eralize to unseen inputs, producing sig- nificantly better results than a dictionary- based baseline.