Splintering Nonconcatenative Languages for Better Tokenization

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Gazit, Bar, Shmidman, Shaltiel, Shmidman, Avi, Pinter, Yuval
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910981822611456
author Gazit, Bar
Shmidman, Shaltiel
Shmidman, Avi
Pinter, Yuval
author_facet Gazit, Bar
Shmidman, Shaltiel
Shmidman, Avi
Pinter, Yuval
contents Common subword tokenization algorithms like BPE and UnigramLM assume that text can be split into meaningful units by concatenative measures alone. This is not true for languages such as Hebrew and Arabic, where morphology is encoded in root-template patterns, or Malay and Georgian, where split affixes are common. We present SPLINTER, a pre-processing step which rearranges text into a linear form that better represents such nonconcatenative morphologies, enabling meaningful contiguous segments to be found by the tokenizer. We demonstrate SPLINTER's merit using both intrinsic measures evaluating token vocabularies in Hebrew, Arabic, and Malay; as well as on downstream tasks using BERT-architecture models trained for Hebrew.
format Preprint
id arxiv_https___arxiv_org_abs_2503_14433
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Splintering Nonconcatenative Languages for Better Tokenization
Gazit, Bar
Shmidman, Shaltiel
Shmidman, Avi
Pinter, Yuval
Computation and Language
Common subword tokenization algorithms like BPE and UnigramLM assume that text can be split into meaningful units by concatenative measures alone. This is not true for languages such as Hebrew and Arabic, where morphology is encoded in root-template patterns, or Malay and Georgian, where split affixes are common. We present SPLINTER, a pre-processing step which rearranges text into a linear form that better represents such nonconcatenative morphologies, enabling meaningful contiguous segments to be found by the tokenizer. We demonstrate SPLINTER's merit using both intrinsic measures evaluating token vocabularies in Hebrew, Arabic, and Malay; as well as on downstream tasks using BERT-architecture models trained for Hebrew.
title Splintering Nonconcatenative Languages for Better Tokenization
topic Computation and Language
url https://arxiv.org/abs/2503.14433