AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

Improving Discrete Diffusion Unmasking Policies Beyond Explicit Reference Policies

Chunsan Hong, Seonho An, Min-Soo Kim, Jong Chul Ye

ICLR 2026 · Image generation › Fast diffusion sampling · approach: Masked diffusion language model decoding · 13 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use learned unmasking policy instead of max-confidenceLLaDA-8B-Instruct · Sudokuaccuracy: 70.5% -> 81.7%Improved
use learned unmasking policy instead of max-confidenceLLaDA-8B-Instruct · Zebraaccuracy: 33.7% -> 36.2%Improved
use learned Top-K unmasking policy instead of max-confidenceLLaDA-8B-Instruct · GSM8Kaccuracy: 68.4% -> 70.3%Improved
use learned Top-K unmasking policy instead of max-confidenceLLaDA-8B-Instruct · Math500accuracy: 27.2% -> 28.4%Improved
use learned policy instead of margin baselineLLaDA-8B-Instruct · Math500accuracy: 0.284 -> 0.284No clear effect
add learned unmasking policy on top of diffu-GRPO model (vs random)diffu-GRPO LLaDA · GSM8Kaccuracy: 0.638 -> 0.764Improved
add learned unmasking policy on top of diffu-GRPO model (vs max-confidence)diffu-GRPO LLaDA · GSM8Kaccuracy: 0.751 -> 0.764Improved
add divergence regularization to reference policy (vs random init without it, DColT) (ablation)UPO policy · Sudokuaccuracy: at least 2%Improved
add divergence regularization to reference policy (vs random init without it, DColT) (ablation)UPO policy · GSM8Kaccuracy: at least 3%Improved
use max-confidence realization instead of Top-K (ablation)UPO policy · SudokuaccuracyImproved
use Top-K realization instead of max-confidence (ablation)UPO policy · GSM8KaccuracyImproved
add divergence term (CE to g_conf / KL to Top-K) to loss (ablation)UPO policy · Sudoku and GSM8KaccuracyImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy