Morphological Typology in BPE Subword Productivity and Language Modeling

TL;DR

This paper explores how morphological typology influences BPE subword productivity and language modeling performance, revealing that synthetic languages benefit more from BPE tokenization and exhibit greater subword regularity.

Contribution

It provides the first systematic analysis of the relationship between morphological typology and BPE-based language modeling performance.

Findings

Synthetic languages show higher subword regularity.

Languages with synthetic morphology perform better in language modeling tasks.

Typological continuum influences BPE tokenization efficiency.

Abstract

This study investigates the impact of morphological typology on tokenization and language modeling performance. We focus on languages with synthetic and analytical morphological structures and examine their productivity when tokenized using the byte-pair encoding (BPE) algorithm. We compare the performance of models trained with similar amounts of data in different languages. Our experiments reveal that languages with synthetic features exhibit greater subword regularity and productivity with BPE tokenization and achieve better results in language modeling tasks. We also observe that the typological continuum from linguistic theory is reflected in several experiments. These findings suggest a correlation between morphological typology and BPE tokenization efficiency.