AIarXiv
Ranked by formulaVerdict: NotableLook Only When Needed: Certificate-Guided Visual Routing for Efficient Diffusion VLMs
Diffusion vision-language models (DVLMs) process their entire input at every denoising step.
Key numbers
- 97% of this input on
- 22.8% of steps on average
- 99.6% of LLaDA-V
Caveats
- Preprint; not yet peer reviewed.
VerdictWorth a reader's time today.
Abstract
Diffusion vision-language models (DVLMs) process their entire input at every denoising step. Since visual tokens make up 89-97% of this input on the benchmarks we study, most of the decoding computation is spent repeatedly processing the same image. Visual token pruning methods partially relieve this burden by removing part of the visual tokens, but they lose the removed information and still process the remaining image at every step. To measure how often the image is actually needed, we compare LLaDA-V's predictions with and without the image at every decoding step and find that removing the image changes a token being unmasked at only 22.8% of steps on average and at as few as 7.9%. We therefore propose LOWN, a decoding method that decides at every step whether to process the image and how many tokens to unmask. LOWN predicts the masked tokens without the image and processes the image only when this prediction disagrees with the last image-conditioned prediction. Agreement certifies the earlier prediction for reuse, and the number of certified tokens sets how many are unmasked in that step. We further observe that the need for the image differs across benchmarks, so LOWN's adaptive decision is needed rather than a fixed schedule. Without training or a tuned threshold, LOWN retains 99.6% of LLaDA-V's score on ten benchmarks at 35.3% of the decoding FLOPs on average. LOWN also generalizes across LLaDA-based DVLMs, composes orthogonally with visual token pruning, and runs under key-value caching for further efficiency gains. Code and project page are available at https://hyeon9698.github.io/LOWN/
The editor's rubric
- Leverage4 / 5
- A general-purpose tool used across several fields (Adam, ResNet, LoRA, next-generation sequencing).
- Magnitude2 / 5
- Solid incremental gain on a meaningful problem.
- Evidence3 / 5
- Solid: multiple benchmarks or cohorts, ablations, fair baselines, released code or data.
- Novelty2 / 5
- A new combination of known ideas.
- Trajectory3 / 5
- A clear path to scale.
- Stakes2 / 5
- Benefits a professional community (practitioners, clinicians, engineers).
Editor’s rationale
Heuristic triage from title and abstract text only, not a reading of the paper. Cues found: breadth (generalizes, many tasks); novelty (alternative to status quo); verification (multiple benchmarks, code released); scale (efficient); stakes (general AI). Red flags: derivative (comparative study).
How the score was computed
Desks place stronger verdicts first, then use this score to order work within the same verdict.
- Merit
- 5.6 / 10
- Adjusted merit
- 4.7 / 10
- Attention
- 86%
- Freshness
- 40%
- Citations0 (reference 15, via semantic-scholar, Oct 30, 2026, 02:05 UTC)
- Influential citations0 (reference 3, via semantic-scholar, Oct 30, 2026, 02:05 UTC)
- Hugging Face upvotes80 (reference 25, via hf-daily, Oct 30, 2026, 13:37 UTC)
- GitHub stars0 (reference 250, via hf-daily, Oct 30, 2026, 13:37 UTC)