The IMS-CUBoulder System for the SIGMORPHON 2020 Shared Task on Unsupervised Morphological Paradigm Completion

In this paper, we present the systems of the University of Stuttgart IMS and\nthe University of Colorado Boulder (IMS-CUBoulder) for SIGMORPHON 2020 Task 2\non unsupervised morphological paradigm completion (Kann et al., 2020). The task\nconsists of generating the morphological paradigms of a set of lemmas, given\nonly the lemmas themselves and unlabeled text. Our proposed system is a\nmodified version of the baseline introduced together with the task. In\nparticular, we experiment with substituting the inflection generation component\nwith an LSTM sequence-to-sequence model and an LSTM pointer-generator network.\nOur pointer-generator system obtains the best score of all seven submitted\nsystems on average over all languages, and outperforms the official baseline,\nwhich was best overall, on Bulgarian and Kannada.\n