Saved in:
Bibliographic Details
Main Authors: Xu, Xianglong, Bowen, John, Taheri, Rojin
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2505.11746
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916741929500672
author Xu, Xianglong
Bowen, John
Taheri, Rojin
author_facet Xu, Xianglong
Bowen, John
Taheri, Rojin
contents While transformer-based models achieve strong performance on text classification, we explore whether masking input tokens can further enhance their effectiveness. We propose token masking regularization, a simple yet theoretically motivated method that randomly replaces input tokens with a special [MASK] token at probability p. This introduces stochastic perturbations during training, leading to implicit gradient averaging that encourages the model to capture deeper inter-token dependencies. Experiments on language identification and sentiment analysis -- across diverse models (mBERT, Qwen2.5-0.5B, TinyLlama-1.1B) -- show consistent improvements over standard regularization techniques. We identify task-specific optimal masking rates, with p = 0.1 as a strong general default. We attribute the gains to two key effects: (1) input perturbation reduces overfitting, and (2) gradient-level smoothing acts as implicit ensembling.
format Preprint
id arxiv_https___arxiv_org_abs_2505_11746
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Token Masking Improves Transformer-Based Text Classification
Xu, Xianglong
Bowen, John
Taheri, Rojin
Computation and Language
Artificial Intelligence
While transformer-based models achieve strong performance on text classification, we explore whether masking input tokens can further enhance their effectiveness. We propose token masking regularization, a simple yet theoretically motivated method that randomly replaces input tokens with a special [MASK] token at probability p. This introduces stochastic perturbations during training, leading to implicit gradient averaging that encourages the model to capture deeper inter-token dependencies. Experiments on language identification and sentiment analysis -- across diverse models (mBERT, Qwen2.5-0.5B, TinyLlama-1.1B) -- show consistent improvements over standard regularization techniques. We identify task-specific optimal masking rates, with p = 0.1 as a strong general default. We attribute the gains to two key effects: (1) input perturbation reduces overfitting, and (2) gradient-level smoothing acts as implicit ensembling.
title Token Masking Improves Transformer-Based Text Classification
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2505.11746