Tackling the Low-resource Challenge for Canonical Segmentation
This work addresses the challenge of morphological segmentation for low-resource languages, which is incremental as it builds on existing methods with specific improvements.
The paper tackled the problem of canonical morphological segmentation in low-resource languages by comparing new models like LSTM pointer-generator and sequence-to-sequence with hard monotonic attention against existing methods, finding that the novel approaches outperformed existing ones by up to 11.4% accuracy in simulated settings, but achieved only 37.4% and 28.4% accuracy for truly low-resource languages Popoluca and Tepehua.
Canonical morphological segmentation consists of dividing words into their standardized morphemes. Here, we are interested in approaches for the task when training data is limited. We compare model performance in a simulated low-resource setting for the high-resource languages German, English, and Indonesian to experiments on new datasets for the truly low-resource languages Popoluca and Tepehua. We explore two new models for the task, borrowing from the closely related area of morphological generation: an LSTM pointer-generator and a sequence-to-sequence model with hard monotonic attention trained with imitation learning. We find that, in the low-resource setting, the novel approaches outperform existing ones on all languages by up to 11.4% accuracy. However, while accuracy in emulated low-resource scenarios is over 50% for all languages, for the truly low-resource languages Popoluca and Tepehua, our best model only obtains 37.4% and 28.4% accuracy, respectively. Thus, we conclude that canonical segmentation is still a challenging task for low-resource languages.