N-gram Prediction and Word Difference Representations for Language Modeling

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Heo, DongNyeong, Rim, Daniela Noemi, Choi, Heeyoul
Format: Preprint
Publié: 2024
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866916382973624320
author Heo, DongNyeong
Rim, Daniela Noemi
Choi, Heeyoul
author_facet Heo, DongNyeong
Rim, Daniela Noemi
Choi, Heeyoul
contents Causal language modeling (CLM) serves as the foundational framework underpinning remarkable successes of recent large language models (LLMs). Despite its success, the training approach for next word prediction poses a potential risk of causing the model to overly focus on local dependencies within a sentence. While prior studies have been introduced to predict future N words simultaneously, they were primarily applied to tasks such as masked language modeling (MLM) and neural machine translation (NMT). In this study, we introduce a simple N-gram prediction framework for the CLM task. Moreover, we introduce word difference representation (WDR) as a surrogate and contextualized target representation during model training on the basis of N-gram prediction framework. To further enhance the quality of next word prediction, we propose an ensemble method that incorporates the future N words' prediction results. Empirical evaluations across multiple benchmark datasets encompassing CLM and NMT tasks demonstrate the significant advantages of our proposed methods over the conventional CLM.
format Preprint
id arxiv_https___arxiv_org_abs_2409_03295
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle N-gram Prediction and Word Difference Representations for Language Modeling
Heo, DongNyeong
Rim, Daniela Noemi
Choi, Heeyoul
Computation and Language
Artificial Intelligence
Causal language modeling (CLM) serves as the foundational framework underpinning remarkable successes of recent large language models (LLMs). Despite its success, the training approach for next word prediction poses a potential risk of causing the model to overly focus on local dependencies within a sentence. While prior studies have been introduced to predict future N words simultaneously, they were primarily applied to tasks such as masked language modeling (MLM) and neural machine translation (NMT). In this study, we introduce a simple N-gram prediction framework for the CLM task. Moreover, we introduce word difference representation (WDR) as a surrogate and contextualized target representation during model training on the basis of N-gram prediction framework. To further enhance the quality of next word prediction, we propose an ensemble method that incorporates the future N words' prediction results. Empirical evaluations across multiple benchmark datasets encompassing CLM and NMT tasks demonstrate the significant advantages of our proposed methods over the conventional CLM.
title N-gram Prediction and Word Difference Representations for Language Modeling
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2409.03295