mlsanityData to 7 Oct 2026

Efficient Diffusion LLMs via Temporal-Spatial Parallel Decoding and Confidence Extrapolation

arXiv 2605.30753 ↗First posted 29 May 2026cs.CLcited by 1

On the map: Efficient LLM inference and serving → Diffusion LLM and MoE inference

Abstract (arXiv)

Diffusion-based large language models (dLLMs) support parallel text generation via iterative denoising, yet inference remains latency-heavy because many steps are spent on redundant refinement and repeated remasking of tokens whose final values are already determined. Prior acceleration methods mainly depend on step-local confidence heuristics or fixed schedules, which are sensitive to prompt and task variation and ignore strong positional effects within a sequence. We cast diffusion decoding as a dynamic control problem and show that token-wise denoising trajectories provide the key signal for reliable control. We propose a trace-aware decoding framework with two components. First, Temporal-Spatial Parallel Decoding (TSPD) uses a lightweight temporalspatial controller that consumes per-token trajectory features, including confidence, entropy, and momentum, together with token position, to decide when a token has converged and can be safely fixed. Second, we introduce Confidence Extrapolation (CE), a training-free state-space module that forecasts future logit trends with uncertainty to support proactive decisions, including safe look-ahead and targeted stabilization when trajectories are oscillatory or underconfident. Together, TSPD and CE reduce unnecessary denoising iterations while preserving output quality, and they compose cleanly with system optimizations such as KV caching.

Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Placed in Diffusion LLM and MoE inference, one of 12 directions in Efficient LLM inference and serving. Builds on 45 of its 51 references that are in the map. First posted on arXiv 29 May 2026.

Builds on

45 of 51 references.

Closest ancestors · 7

Same direction, newest first

  1. 2025Accelerating Diffusion LLM Inference via Local Determinism Propagation8 Oct
  2. 2025CreditDecoding: Accelerating Parallel Decoding in Diffusion Large Language Models with Trace Credit7 Oct
  3. 2025dParallel: Learnable Parallel Decoding for dLLMs30 Sep
  4. 2025Learning to Parallel: Accelerating Diffusion Large Language Models via Learnable Parallel Decoding29 Sep
  5. 2025Diffusion LLMs Can Do Faster-Than-AR Inference via Discrete Diffusion Forcing8 Aug
  6. 2025Accelerating Diffusion LLMs via Adaptive Parallel Decoding31 May
  7. 2025Fast-dLLM: Training-free Acceleration of Diffusion LLM by Enabling KV Cache and Parallel Decoding28 May

Other references · 38

  1. 2020Denoising Diffusion Probabilistic Modelscited by 20,851
  2. 2019Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformercited by 14,491
  3. 2021Training Verifiers to Solve Math Word Problemscited by 9,743
  4. 2020Denoising Diffusion Implicit Modelscited by 9,655
  5. 2021Evaluating Large Language Models Trained on Codecited by 8,053
  6. 2015Deep Unsupervised Learning using Nonequilibrium Thermodynamicscited by 7,005
  7. 2021Measuring Mathematical Problem Solving With the MATH Datasetcited by 5,203
  8. 2019Generative Modeling by Estimating Gradients of the Data Distributioncited by 4,067
  9. 2014Learning Phrase Representations using RNN Encoder-Decoder for Statistical Machine Translationcited by 4,016
  10. 2021Improved Denoising Diffusion Probabilistic Modelscited by 3,736
  11. 2021Program Synthesis with Large Language Modelscited by 3,509
  12. 2021Structured Denoising Diffusion Models in Discrete State-Spacescited by 1,665
26 more
  1. 2025Large Language Diffusion Modelscited by 872
  2. 2024Simple and Effective Masked Diffusion Language Modelscited by 800
  3. 2023Discrete Diffusion Modeling by Estimating the Ratios of the Data Distributioncited by 712
  4. 2021Argmax Flows and Multinomial Diffusion: Learning Categorical Distributionscited by 566
  5. 2022Efficiently Scaling Transformer Inferencecited by 529
  6. 2025Dream 7B: Diffusion Large Language Modelscited by 515
  7. 2022A Continuous Time Framework for Discrete Denoising Modelscited by 437
  8. 2016Bidirectional Attention Flow for Machine Comprehensioncited by 278
  9. 2024Scaling Diffusion Language Models via Adaptation from Autoregressive Modelscited by 268
  10. 2025Mercury: Ultra-Fast Language Models Based on Diffusioncited by 225
  11. 2024Scaling up Masked Diffusion Models on Textcited by 195
  12. 2025d1: Scaling Reasoning in Diffusion Large Language Models via Reinforcement Learningcited by 178
  13. 2025dLLM-Cache: Accelerating Diffusion Large Language Models with Adaptive Cachingcited by 168
  14. 2025Fast-dLLM v2: Efficient Block-Diffusion LLMcited by 128
  15. 2024MiniCache: KV Cache Compression in Depth Dimension for Large Language Modelscited by 122
  16. 2025FlashDLM: Accelerating Diffusion Language Model Inference via Efficient KV Caching and Guided Diffusioncited by 77
  17. 2025Diffusion Language Models Know the Answer Before Decodingcited by 68
  18. 2025LLaDA-MoE: A Sparse MoE Diffusion Language Modelcited by 54
  19. 2025Accelerating Diffusion Large Language Models with SlowFast Sampling: The Three Golden Principlescited by 43
  20. 2025Time Is a Feature: Exploiting Temporal Dynamics in Diffusion Language Modelscited by 33
  21. 2025dInfer: An Efficient Inference Framework for Diffusion Language Modelscited by 33
  22. 2024Autoregressive Large Language Models are Computationally Universalcited by 14
  23. 2025Beyond Confidence: Adaptive and Coherent Decoding for Diffusion Language Modelscited by 9
  24. 2025Decoding Large Language Diffusion Models with Foreseeing Movementcited by 6
  25. 2023Bidirectional Attention as a Mixture of Continuous Word Expertscited by 1
  26. 2025Learning to Look at the Other Side: A Semantic Probing Study of Word Embeddings in LLMs with Enabled Bidirectional Attentioncited by 1

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026CORA-Diff: Confidence-Oriented Residual Acceptance for Efficient Diffusion Language Model Inferencecited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026Trajectory-Level Speculative Decoding for Diffusion Language Models
  2. 2026Roll Out and Roll Back: Diffusion LLMs are Their Own Efficiency Teachers
  3. 2026PSD: Pushing the Pareto Frontier of Diffusion LLMs via Parallel Speculative Decoding
  4. 2026Early-Bird Decoding: Accelerating Diffusion LLMs with Learnable Block Sizes and Parallel Sampling
  5. 2026Stability-Weighted Decoding for Diffusion Language Models
  6. 2026Dynamic-dLLM: Dynamic Cache-Budget and Adaptive Parallel Decoding for Training-Free Acceleration of Diffusion LLM
  7. 2025From Bits to Rounds: Parallel Decoding with Exploration for Diffusion Language Models
  8. 2026Mean-Field Parallel Decoding for Discrete Diffusion Language Models