Bootstrapping Techniques for Polysynthetic Morphological Analysis
This addresses the problem of limited NLP tools for polysynthetic languages, offering a practical solution for linguists and communities, though it is incremental in applying known bootstrapping techniques to a specific domain.
The paper tackles the challenge of developing natural language technologies for polysynthetic languages with large, sparse vocabularies and scarce data by bootstrapping a neural morphological analyzer for Kunwinjku, achieving 94.7% accuracy and a 10 percentage point improvement over the baseline.
Polysynthetic languages have exceptionally large and sparse vocabularies, thanks to the number of morpheme slots and combinations in a word. This complexity, together with a general scarcity of written data, poses a challenge to the development of natural language technologies. To address this challenge, we offer linguistically-informed approaches for bootstrapping a neural morphological analyzer, and demonstrate its application to Kunwinjku, a polysynthetic Australian language. We generate data from a finite state transducer to train an encoder-decoder model. We improve the model by "hallucinating" missing linguistic structure into the training data, and by resampling from a Zipf distribution to simulate a more natural distribution of morphemes. The best model accounts for all instances of reduplication in the test set and achieves an accuracy of 94.7% overall, a 10 percentage point improvement over the FST baseline. This process demonstrates the feasibility of bootstrapping a neural morph analyzer from minimal resources.