Work

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

Year: 2026

Type: conference-paper

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.

Source: International Joint Conference on Artificial Intelligence

DOI: 10.24963/ijcai.2026/873

Authors: Daehoon Gwak, Minhyung Lee, Junwoo Park, Jaegul Choo

Institutions: Korea Advanced Institute of Science and Technology, Yonsei University

Language: English

Topic: Natural Language Processing Techniques

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

Open Access status: bronze

Funders: Korea Advanced Institute of Science and Technology, Ministry of Science and ICT, South Korea, Institute for Information and Communications Technology Promotion

Awards: G3736587842, G4468082379, G8705040809