A natural desideratum for generative models is self-correction--detecting and revising low-quality tokens at inference. While Masked Diffusion Models (MDMs) have emerged as a promising approach for generative modeling in discrete spaces, their capacity for self-correction remains poorly understood. Prior attempts to incorporate self-correction into MDMs either require overhauling MDM architectures/training or rely on imprecise proxies for token quality, limiting their applicability. Motivated by this, we introduce PRISM--Plug-in Remasking for Inference-time Self-correction of Masked Diffusions--a lightweight, model-agnostic approach that applies to any pretrained MDM. Theoretically, PRISM defines a self-correction loss that provably learns per-token quality scores, without RL or a verifier. These quality scores are computed in the same forward pass with MDM and used to detect low-quality tokens. Empirically, PRISM advances MDM inference across domains and scales: Sudoku; unconditional text (170M); and code with LLaDA (8B).
@article{arxiv.2510.01384,
title = {Fine-Tuning Masked Diffusion for Provable Self-Correction},
author = {Jaeyeon Kim and Seunggeun Kim and Taekyun Lee and David Z. Pan and Hyeji Kim and Sham Kakade and Sitan Chen},
journal= {arXiv preprint arXiv:2510.01384},
year = {2026}
}
Comments
Authorship statement: Jaeyeon Kim and Seunggeun Kim contributed equally, and Taekyun Lee is also a co first author