mlsanityData to 14 Oct 2026

From Position Risks to Block Survival: Faster Generation for Diffusion Language Models

arXiv 2609.33390 ↗First posted 27 Sep 2026cs.CL cs.AIcited by 0

On the map: Diffusion models → Discrete and text diffusion

Abstract (arXiv)

Diffusion language models (DLMs) can accelerate generation by predicting multiple tokens in parallel, but there is a mismatch between how these tokens are predicted and how they ultimately contribute to generation. Parallel predictions can hardly condition on the tokens selected earlier within the same block, even though their validity depends on this realized prefix. Under the popular proposal-verification decoding, this mismatch makes errors highly asymmetric: an early rejection prevents all subsequent proposals from contributing decoding progress. We introduce BRISK-DLM, a framework that addresses both mismatches by optimizing proposal learning and selection for verified progress. BRISK-DLM trains on self-generated sequences, using risk-reward weighting to dynamically prioritize positions by their impact on verified progress and decoding cost. During inference, a lightweight prefix-conditioned corrector reranks existing candidates using previously selected tokens and preferences distilled from the model's own verifier. The corrector reuses the backbone's parallel representations and requires no additional backbone evaluation, while fused execution keeps its overhead small. BRISK-DLM improves end-to-end throughput by up to 37.4% while preserving task quality, establishing a new quality-throughput frontier for DLM generation.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Discrete and text diffusion, one of 14 directions in Diffusion models. Builds on 29 of its 32 references that are in the map. First posted on arXiv 27 Sep 2026.

Builds on

29 of 32 references.

Closest ancestors · 15

Same direction, newest first

  1. 2026Data-Efficient Autoregressive-to-Diffusion Language Models via On-Policy Distillation4 Jun
  2. 2026Introspective Diffusion Language Models13 Apr
  3. 2026S2D2: Fast Decoding for Diffusion LLMs via Training-Free Self-Speculation26 Mar
  4. 2026LLaDA2.1: Speeding Up Text Diffusion via Token Editing9 Feb
  5. 2026DFlash: Block Diffusion for Flash Speculative Decoding5 Feb
  6. 2025Efficient-DLM: From Autoregressive to Diffusion Language Models, and Beyond in Speed16 Dec
  7. 2025LLaDA2.0: Scaling Up Diffusion Language Models to 100B10 Dec
  8. 2025SDAR: A Synergistic Diffusion-AutoRegression Paradigm for Scalable Sequence Generation7 Oct
  9. 2025Dream 7B: Diffusion Large Language Models21 Aug
  10. 2025Block Diffusion: Interpolating Between Autoregressive and Diffusion Language Models12 Mar
  11. 2025Large Language Diffusion Models14 Feb
  12. 2024Beyond Autoregression: Fast LLMs via Self-Distillation Through Time28 Oct
3 more
  1. 2024Scaling Diffusion Language Models via Adaptation from Autoregressive Models23 Oct
  2. 2024Simple and Effective Masked Diffusion Language Models11 Jun
  3. 2023Discrete Diffusion Modeling by Estimating the Ratios of the Data Distribution25 Oct

Other references · 14

  1. 2021Structured Denoising Diffusion Models in Discrete State-Spacescited by 1,719
  2. 2022Fast Inference from Transformers via Speculative Decodingcited by 1,653
  3. 2022Diffusion-LM Improves Controllable Text Generationcited by 1,027
  4. 2023Accelerating Large Language Model Decoding with Speculative Samplingcited by 1,006
  5. 2024Medusa: Simple LLM Inference Acceleration Framework with Multiple Decoding Headscited by 843
  6. 2024EAGLE: Speculative Sampling Requires Rethinking Feature Uncertaintycited by 578
  7. 2025Fast-dLLM: Training-free Acceleration of Diffusion LLM by Enabling KV Cache and Parallel Decodingcited by 420
  8. 2024Better & Faster Large Language Models via Multi-token Predictioncited by 360
  9. 2025EAGLE-3: Scaling up Inference Acceleration of Large Language Models via Training-Time Testcited by 331
  10. 2026DSpark: Confidence-Scheduled Speculative Decoding with Semi-Autoregressive Generationcited by 45
  11. 2026Domino: Decoupling Causal Modeling from Autoregressive Drafting in Speculative Decodingcited by 34
  12. 2025Free Draft-and-Verification: Toward Lossless Parallel Decoding for Diffusion Large Language Modelscited by 19
2 more
  1. 2026D-PACE: Dynamic Position-Aware Cross-Entropy for Parallel Speculative Draftingcited by 17
  2. 2025GRIFFIN: Effective Token Alignment for Faster Speculative Decodingcited by 16

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026Diffusion Language Models: An Experimental Analysis
  2. 2025On Powerful Ways to Generate: Autoregression, Diffusion, and Beyond
  3. 2026Accelerating Masked Diffusion Large Language Models: A Survey of Efficient Inference Techniques
  4. 2025From Next-Token to Next-Block: A Principled Adaptation Path for Diffusion LLMs
  5. 2025A Survey on Diffusion Language Models
  6. 2025Sequential Diffusion Language Models
  7. 2025Diffusion Language Models are Provably Optimal Parallel Samplers
  8. 2025Diffusion Language Model Inference with Monte Carlo Tree Search