Unsupervised Morphological Tree Tokenizer

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zhu, Qingyang, Hu, Xiang, Ji, Pengyu, Wu, Wei, Tu, Kewei
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913935407448064
author Zhu, Qingyang
Hu, Xiang
Ji, Pengyu
Wu, Wei
Tu, Kewei
author_facet Zhu, Qingyang
Hu, Xiang
Ji, Pengyu
Wu, Wei
Tu, Kewei
contents As a cornerstone in language modeling, tokenization involves segmenting text inputs into pre-defined atomic units. Conventional statistical tokenizers often disrupt constituent boundaries within words, thereby corrupting semantic information. To address this drawback, we introduce morphological structure guidance to tokenization and propose a deep model to induce character-level structures of words. Specifically, the deep model jointly encodes internal structures and representations of words with a mechanism named $\textit{MorphOverriding}$ to ensure the indecomposability of morphemes. By training the model with self-supervised objectives, our method is capable of inducing character-level structures that align with morphological rules without annotated training data. Based on the induced structures, our algorithm tokenizes words through vocabulary matching in a top-down manner. Empirical results indicate that the proposed method effectively retains complete morphemes and outperforms widely adopted methods such as BPE and WordPiece on both morphological segmentation tasks and language modeling tasks. Code is available at https://github.com/martianmartina/TreeTokenizer.
format Preprint
id arxiv_https___arxiv_org_abs_2406_15245
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Unsupervised Morphological Tree Tokenizer
Zhu, Qingyang
Hu, Xiang
Ji, Pengyu
Wu, Wei
Tu, Kewei
Computation and Language
Machine Learning
As a cornerstone in language modeling, tokenization involves segmenting text inputs into pre-defined atomic units. Conventional statistical tokenizers often disrupt constituent boundaries within words, thereby corrupting semantic information. To address this drawback, we introduce morphological structure guidance to tokenization and propose a deep model to induce character-level structures of words. Specifically, the deep model jointly encodes internal structures and representations of words with a mechanism named $\textit{MorphOverriding}$ to ensure the indecomposability of morphemes. By training the model with self-supervised objectives, our method is capable of inducing character-level structures that align with morphological rules without annotated training data. Based on the induced structures, our algorithm tokenizes words through vocabulary matching in a top-down manner. Empirical results indicate that the proposed method effectively retains complete morphemes and outperforms widely adopted methods such as BPE and WordPiece on both morphological segmentation tasks and language modeling tasks. Code is available at https://github.com/martianmartina/TreeTokenizer.
title Unsupervised Morphological Tree Tokenizer
topic Computation and Language
Machine Learning
url https://arxiv.org/abs/2406.15245