Corpus Frequencies in Morphological Inflection: Do They Matter?

TL;DR

The paper tackles morphological inflection by integrating corpus frequency information into three stages: train-dev-test data splitting, training data sampling, and evaluation. It introduces a frequency-weighted, lemma-disjoint split, a frequency-aware training regime with a tunable corpus-frequency temperature $\tau$, and token accuracy as a deployment-relevant evaluation metric that weights frequent forms. Across 43 UD languages, frequency-aware training yields gains in token accuracy for 26 languages and improves type accuracy in 41 of 43 languages, with the best average observed near $\tau=0.5$. The study provides practical guidance for deployment of inflection systems under real-world word distributions and makes the code publicly available for replication and extension.

Abstract

The traditional approach to morphological inflection (the task of modifying a base word (lemma) to express grammatical categories) has been, for decades, to consider lexical entries of lemma-tag-form triples uniformly, lacking any information about their frequency distribution. However, in production deployment, one might expect the user inputs to reflect a real-world distribution of frequencies in natural texts. With future deployment in mind, we explore the incorporation of corpus frequency information into the task of morphological inflection along three key dimensions during system development: (i) for train-dev-test split, we combine a lemma-disjoint approach, which evaluates the model's generalization capabilities, with a frequency-weighted strategy to better reflect the realistic distribution of items across different frequency bands in training and test sets; (ii) for evaluation, we complement the standard type accuracy (often referred to simply as accuracy), which treats all items equally regardless of frequency, with token accuracy, which assigns greater weight to frequent words and better approximates performance on running text; (iii) for training data sampling, we introduce a method novel in the context of inflection, frequency-aware training, which explicitly incorporates word frequency into the sampling process. We show that frequency-aware training outperforms uniform sampling in 26 out of 43 languages.

Figures (1)

• Figure 1: The morphological inflection task: an input-output example, inflection of English lemma "well" to superlative, inflected form "best" (figure adopted from tsd).