Accelerating Masked Diffusion Large Language Models: A Survey of Efficient Inference Techniques

• URL: http://arxiv.org/abs/2607.12829v1
• Date: Tue, 14 Jul 2026 14:48:06 GMT
• Title: Accelerating Masked Diffusion Large Language Models: A Survey of Efficient Inference Techniques
• Abstract summary: Diffusion large language models (dLLMs) offer a theoretical advantage in parallel generation over standard autoregressive models.<n> realizing this efficiency requires specialized inference mechanisms, such as diffusion-aware caching and reuse.<n>We introduce a unified latency decomposition framework for dLLMs to disentangle these factors and analyze their impact on inference speed in real deployments.
• Score: 45.003333444832656
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Diffusion large language models (dLLMs) offer a theoretical advantage in parallel generation over standard autoregressive models. However, parallel generation alone does not guarantee practical speedups. Realizing this efficiency requires specialized inference mechanisms, such as diffusion-aware caching and reuse. Consequently, as inference efficiency becomes a prerequisite for practical deployment, recent research has actively explored acceleration techniques across algorithms, architectures, and systems. However, rigorous comparisons remain difficult, as end-to-end latency stems from intricate trade-offs between algorithmic, architectural, and system-level factors that are often conflated in existing benchmarks. In this survey, we introduce a unified latency decomposition framework for dLLMs to disentangle these factors and analyze their impact on inference speed in real deployments. Guided by this framework, we categorize acceleration techniques along three axes covering algorithmic innovations, architectural and system optimizations, and inference-time scaling. Finally, we provide guidelines for reproducible benchmarking and highlight open challenges for realizing the full potential of parallel generation.

Related papers

• FLARE: Diffusion for Hybrid Language Model [72.60770374799634]
  FLARE is a systematic conversion framework for hybrid-attention large language models.<n>It combines a token-equal AR-and-diffusion objective, hardware-aware kernels, and unified inference, enabling one checkpoint to support both AR-style verified decoding and diffusion-style parallel denoising.<n>Our results suggest practical dLLMs are limited not only by decoding algorithms, but also by transfer data quality and the training inefficiency of current block-diffusion objectives.
  arXiv  Detail & Related papers  (2026-06-01T06:58:15Z)
• Deep Unfolding: Recent Developments, Theory, and Design Guidelines [99.63555420898554]
  This article provides a tutorial-style overview of deep unfolding, a framework that transforms optimization algorithms into structured, trainable ML architectures.<n>We review the foundations of optimization for inference and for learning, introduce four representative design paradigms for deep unfolding, and discuss the distinctive training schemes that arise from their iterative nature.
  arXiv  Detail & Related papers  (2025-12-03T13:16:35Z)
• Beyond Confidence: Adaptive and Coherent Decoding for Diffusion Language Models [64.92045568376705]
  Coherent Contextual Decoding (CCD) is a novel inference framework built upon two core innovations.<n>CCD employs a trajectory rectification mechanism that leverages historical context to enhance sequence coherence.<n>Instead of rigid allocations based on diffusion steps, we introduce an adaptive sampling strategy that dynamically adjusts the unmasking budget for each step.
  arXiv  Detail & Related papers  (2025-11-26T09:49:48Z)
• A Survey on Cache Methods in Diffusion Models: Toward Efficient Multi-Modal Generation [15.689880312464004]
  Diffusion Models have become a cornerstone of modern generative AI for their exceptional generation quality and controllability.<n>Diffusion Caching offers a training-free, architecture-agnostic, and efficient inference paradigm.<n>By enabling computation feature-level cross-step reuse and inter-layer scheduling, it reduces without modifying model parameters.
  arXiv  Detail & Related papers  (2025-10-22T16:46:05Z)
• How Efficient Are Diffusion Language Models? A Critical Examination of Efficiency Evaluation Practices [81.85465545346266]
  Diffusion language models (DLMs) have emerged as a promising alternative to the long-dominant autoregressive (AR) paradigm.<n>Yet, current open-source DLMs often underperform their AR counterparts in speed, limiting their real-world utility.<n>This work presents a systematic study of DLM efficiency, identifying key issues in prior evaluation methods.
  arXiv  Detail & Related papers  (2025-10-21T10:00:32Z)
• READER: Retrieval-Assisted Drafter for Efficient LLM Inference [0.0386965802948046]
  Autoregressive Language Models instantiate a factorized likelihood over token sequences, yet their strictly sequential decoding process imposes an intrinsic lower bound on latency inference.<n>This bottleneck has emerged as a central obstacle to the scalable deployment of large-scale generative models.<n>We present READER, a speculative decoding framework that bypasses the training of the auxiliary draft model.
  arXiv  Detail & Related papers  (2025-08-12T16:47:48Z)
• Inference Acceleration of Autoregressive Normalizing Flows by Selective Jacobi Decoding [12.338918067455436]
  Normalizing flows are promising generative models with advantages such as theoretical rigor, analytical log-likelihood, and end-to-end training.<n>Recent advancements utilize autoregressive modeling, significantly enhancing expressive power and generation quality.<n>We propose a selective Jacobi decoding (SeJD) strategy that accelerates autoregressive inference through parallel iterative optimization.
  arXiv  Detail & Related papers  (2025-05-30T16:53:15Z)
• Faster Diffusion Action Segmentation [9.868244939496678]
  Temporal Action Classification (TAS) is an essential task in video analysis, aiming to segment and classify continuous frames into distinct action segments. Recent advances in diffusion models have demonstrated substantial success in TAS tasks due to their stable training process and high-quality generation capabilities. We propose EffiDiffAct, an efficient and high-performance TAS algorithm.
  arXiv  Detail & Related papers  (2024-08-04T13:23:18Z)
• Inference Optimization of Foundation Models on AI Accelerators [68.24450520773688]
  Powerful foundation models, including large language models (LLMs), with Transformer architectures have ushered in a new era of Generative AI. As the number of model parameters reaches to hundreds of billions, their deployment incurs prohibitive inference costs and high latency in real-world scenarios. This tutorial offers a comprehensive discussion on complementary inference optimization techniques using AI accelerators.
  arXiv  Detail & Related papers  (2024-07-12T09:24:34Z)
• Consistency Models for Scalable and Fast Simulation-Based Inference [9.27488642055461]
  We present consistency models for posterior estimation (CMPE), a new conditional sampler for simulation-based inference ( SBI) CMPE essentially distills a continuous probability flow and enables rapid few-shot inference with an unconstrained architecture. Our empirical evaluation demonstrates that CMPE not only outperforms current state-of-the-art algorithms on hard low-dimensional benchmarks, but also achieves competitive performance with much faster sampling speed.
  arXiv  Detail & Related papers  (2023-12-09T02:14:12Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.