Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Zhang, Yang, Bartley, Travis M., Graterol-Fuenmayor, Mariana, Lavrukhin, Vitaly, Bakhturina, Evelina, Ginsburg, Boris
Format: Preprint
Veröffentlicht: 2023
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2309.13426
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866913198448312320
author Zhang, Yang
Bartley, Travis M.
Graterol-Fuenmayor, Mariana
Lavrukhin, Vitaly
Bakhturina, Evelina
Ginsburg, Boris
author_facet Zhang, Yang
Bartley, Travis M.
Graterol-Fuenmayor, Mariana
Lavrukhin, Vitaly
Bakhturina, Evelina
Ginsburg, Boris
contents Text normalization - the conversion of text from written to spoken form - is traditionally assumed to be an ill-formed task for language models. In this work, we argue otherwise. We empirically show the capacity of Large-Language Models (LLM) for text normalization in few-shot scenarios. Combining self-consistency reasoning with linguistic-informed prompt engineering, we find LLM based text normalization to achieve error rates around 40\% lower than top normalization systems. Further, upon error analysis, we note key limitations in the conventional design of text normalization tasks. We create a new taxonomy of text normalization errors and apply it to results from GPT-3.5-Turbo and GPT-4.0. Through this new framework, we can identify strengths and weaknesses of GPT-based TN, opening opportunities for future work.
format Preprint
id arxiv_https___arxiv_org_abs_2309_13426
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle A Chat About Boring Problems: Studying GPT-based text normalization
Zhang, Yang
Bartley, Travis M.
Graterol-Fuenmayor, Mariana
Lavrukhin, Vitaly
Bakhturina, Evelina
Ginsburg, Boris
Computation and Language
Artificial Intelligence
Text normalization - the conversion of text from written to spoken form - is traditionally assumed to be an ill-formed task for language models. In this work, we argue otherwise. We empirically show the capacity of Large-Language Models (LLM) for text normalization in few-shot scenarios. Combining self-consistency reasoning with linguistic-informed prompt engineering, we find LLM based text normalization to achieve error rates around 40\% lower than top normalization systems. Further, upon error analysis, we note key limitations in the conventional design of text normalization tasks. We create a new taxonomy of text normalization errors and apply it to results from GPT-3.5-Turbo and GPT-4.0. Through this new framework, we can identify strengths and weaknesses of GPT-based TN, opening opportunities for future work.
title A Chat About Boring Problems: Studying GPT-based text normalization
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2309.13426