Evaluating Neural Morphological Taggers for Sanskrit

TL;DR

This paper evaluates four neural sequence labelling models for Sanskrit morphological tagging, highlighting the importance of models that explicitly handle label structure due to the language's complex morphology.

Contribution

It provides a comparative analysis of neural models on Sanskrit, emphasizing the need for structure-aware models for large label spaces.

Findings

Explicit label structure modeling improves generalization.

Syncretism causes common errors across models.

Some neural models outperform others on Sanskrit tagging.

Abstract

Neural sequence labelling approaches have achieved state of the art results in morphological tagging. We evaluate the efficacy of four standard sequence labelling models on Sanskrit, a morphologically rich, fusional Indian language. As its label space can theoretically contain more than 40,000 labels, systems that explicitly model the internal structure of a label are more suited for the task, because of their ability to generalise to labels not seen during training. We find that although some neural models perform better than others, one of the common causes for error for all of these models is mispredictions due to syncretism.