New Method Accelerates Diffusion Model Inference Through Block Verification
A new technique called block verification adapts speculative decoding from large language models to continuous diffusion models, achieving up to a 6.3% speedup over existing speculative methods. Speculative decoding accelerates inference by using a smaller draft model whose outputs are accepted or rejected to match a target distribution, but applying this to continuous diffusion models has been technically challenging due to difficulties sampling from residual distributions in continuous space. The work also formalizes a training-free 'Free Drafter' heuristic, potentially lowering the barrier to faster diffusion inference without additional computational cost.
Researchers have introduced a novel speculative sampling scheme designed to accelerate inference in continuous diffusion models, a class of generative models widely used for image and audio synthesis. Speculative decoding, originally developed for large language models (LLMs), uses a lightweight draft model to propose outputs that are then verified against a more expensive target model, preserving the target distribution while reducing computation. Extending this to diffusion models has proven difficult because the residual distribution sampling required in continuous spaces is non-trivial, leading prior approaches to use computationally inefficient methods or alternative schemes that sacrifice theoretical guarantees. The proposed method enables block verification — a technique from LLMs that verifies multiple draft steps simultaneously and provably improves acceptance rates — to be applied to diffusion models for the first time. Additionally, the paper formalizes the 'Free Drafter,' a self-speculative, training-free heuristic that, combined with block verification, yields up to a 6.3% speedup over existing speculative diffusion methods with negligible overhead. The work was submitted to arXiv in June 2026 and has not yet undergone formal peer review.
What's missing
The paper has not yet been peer-reviewed. Key open questions include: how the speedup scales across different diffusion model architectures and modalities (e.g., video, audio) beyond those tested; whether the 6.3% speedup holds under varying hardware configurations; and how the Free Drafter's performance compares to trained draft models in absolute quality terms. The paper's own scope is limited to the specific models and benchmarks evaluated, and broader generalization remains to be demonstrated.
What different sources said
- arXiv stat.MLCenter
Accelerating Speculative Diffusions via Block Verification
Related
Gut Bacteria Enzyme Found to Break Down Heat-Processed Food Compounds, Producing Novel Biogenic Amines
Researchers have discovered that an enzyme in common gut bacteria can degrade N-epsilon-carboxymethyllysine (CML), a compound formed during thermal food processing, producing previously unknown biogenic amines. The enzyme, ornithine decarboxylase SpeC from enterobacteria, acts on CML and related modified lysine derivatives through a low-level 'underground' catalytic activity. This finding suggests a previously unrecognized communication axis between thermally processed dietary compounds and gut microbial physiology, with potential implications for host health.
Full-Length Gene Sequencing Reveals Two Distinct Bacterial Communities in Black-Legged Ticks Expanding Into Canada
Researchers used Oxford Nanopore full-length 16S rRNA gene sequencing to characterize the microbiome of Ixodes scapularis black-legged ticks collected in Nova Scotia, Canada, distinguishing between tick-adapted bacteria and environmentally acquired bacteria. The study comes as I. scapularis — the primary vector of Lyme disease — is rapidly expanding northward into Canada due to climate change. The findings suggest that environmentally derived bacteria in tick microbiomes are not mere contamination, which has implications for how tick microbiome data is collected and interpreted across surveillance studies.
Study Identifies Metabolic Link Between Cell Envelope Stress and Biofilm Formation in Bacteria
Researchers have discovered that the metabolite acetyl-CoA directly inhibits enzymes that degrade the bacterial signaling molecule c-di-GMP, connecting cell envelope biosynthesis stress to biofilm formation in Pseudomonas aeruginosa. The study found that sub-inhibitory concentrations of antibiotics targeting early peptidoglycan biosynthesis — but not other antibiotic classes — elevate c-di-GMP levels by reducing phosphodiesterase activity, with acetyl-CoA competing for the enzyme active site. Because the relevant enzyme domain is broadly conserved across bacterial species, this checkpoint mechanism may be widespread and could have implications for understanding antibiotic-induced biofilm responses.