Where-to-Unmask: Ground-Truth-Guided Unmasking Order Learning for Masked Diffusion Language Models

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Asano, Hikaru, Kozuno, Tadashi, Saito, Kuniaki, Baba, Yukino
Format: Preprint
Publié: 2026
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866912894054039552
author Asano, Hikaru
Kozuno, Tadashi
Saito, Kuniaki
Baba, Yukino
author_facet Asano, Hikaru
Kozuno, Tadashi
Saito, Kuniaki
Baba, Yukino
contents Masked Diffusion Language Models (MDLMs) generate text by iteratively filling masked tokens, requiring two coupled decisions at each step: which positions to unmask (where-to-unmask) and which tokens to place (what-to-unmask). While standard MDLM training directly optimizes token prediction (what-to-unmask), inference-time unmasking orders (where-to-unmask) are typically determined by heuristic confidence measures or trained through reinforcement learning with costly on-policy rollouts. To address this, we introduce Gt-Margin, a position-wise score derived from ground-truth tokens, defined as the probability margin between the correct token and its strongest alternative. Gt-Margin yields an oracle unmasking order that prioritizes easier positions first under each partially masked state. We demonstrate that leveraging this oracle unmasking order significantly enhances final generation quality, particularly on logical reasoning benchmarks. Building on this insight, we train a supervised unmasking planner via learning-to-rank to imitate the oracle ordering from masked contexts. The resulting planner integrates into standard MDLM sampling to select where-to-unmask, improving reasoning accuracy without modifying the token prediction model.
format Preprint
id arxiv_https___arxiv_org_abs_2602_09501
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Where-to-Unmask: Ground-Truth-Guided Unmasking Order Learning for Masked Diffusion Language Models
Asano, Hikaru
Kozuno, Tadashi
Saito, Kuniaki
Baba, Yukino
Computation and Language
Masked Diffusion Language Models (MDLMs) generate text by iteratively filling masked tokens, requiring two coupled decisions at each step: which positions to unmask (where-to-unmask) and which tokens to place (what-to-unmask). While standard MDLM training directly optimizes token prediction (what-to-unmask), inference-time unmasking orders (where-to-unmask) are typically determined by heuristic confidence measures or trained through reinforcement learning with costly on-policy rollouts. To address this, we introduce Gt-Margin, a position-wise score derived from ground-truth tokens, defined as the probability margin between the correct token and its strongest alternative. Gt-Margin yields an oracle unmasking order that prioritizes easier positions first under each partially masked state. We demonstrate that leveraging this oracle unmasking order significantly enhances final generation quality, particularly on logical reasoning benchmarks. Building on this insight, we train a supervised unmasking planner via learning-to-rank to imitate the oracle ordering from masked contexts. The resulting planner integrates into standard MDLM sampling to select where-to-unmask, improving reasoning accuracy without modifying the token prediction model.
title Where-to-Unmask: Ground-Truth-Guided Unmasking Order Learning for Masked Diffusion Language Models
topic Computation and Language
url https://arxiv.org/abs/2602.09501