AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

Trajectory-Level Speculative Decoding for Diffusion Language Models

Tianxiang Pan, Baitao Gong, Mo Guang, Hongwei Yong, Tianpeng Jiang, Yaqian Li, Zheng Cao, Kaiwen Long

ICML 2026 · Image generation › Fast diffusion sampling · approach: Masked diffusion language model decoding · 0 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
add trajectory-level speculative decodingFast-dLLM (DC) · GSM8K (LLaDA)denoising steps: 110.3 -> 71.3Improved
add trajectory-level speculative decodingFast-dLLM (DC) · HumanEval (LLaDA)denoising steps: 179.2 -> 106.7Improved
add trajectory-level speculative decodingFast-dLLM (DC) · HumanEval (LLaDA)TPS: 70.6 -> 94.4Improved
add trajectory-level speculative decodingFast-dLLM (DC) · GSM8K (Dream)accuracy: 74.6 -> 73.7Hurt
add trajectory-level speculative decodingFast-dLLM (DC) · HumanEval (Dream)accuracy: 51.2 -> 54.3Improved
add trajectory-level speculative decodingFast-dLLM (DC) · MATH level 4 (LLaDA)accuracy: 30.4 -> 25.1Hurt
add trajectory-level speculative decodingFast-dLLM (DC) · IFEval (LLaDA, H800)TPS: 36.8 -> 69.1Improved
add intra-block tree W1D1(1) (ablation)Fast-dLLM (DC) · HumanEval (LLaDA)tokens per step: 2.61 -> 3.51Improved
grow intra-block tree from W2D2(3) to W3D3(6) (ablation)trajectory speculation · HumanEval (LLaDA)latency (s): 5.3 -> 5.5Hurt
add inter-block speculation to W2D2(3) (ablation)trajectory speculation · HumanEval (LLaDA)denoising steps: 124.0 -> 106.7Improved
add inter-block speculation to W2D2(3) (ablation)trajectory speculation · HumanEval (LLaDA)latency (s): 5.3 -> 5.0Improved
use full tree W3D3(9) instead of hybrid W3D3(6) (ablation)trajectory speculation · HumanEval (LLaDA)total latency (s): 5.5 -> 6.6Hurt

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