/ Paper arXiv 2611.08449
Curator's notes
VFlash jointly decodes visual primitives for fast, accurate planning, achieving ~25× speedup and higher task success.
VFlash predicts whole shapes at once, making planning much faster and more accurate for robots and cars.VFlash jointly decodes visual primitives for fast, accurate planning, achieving ~25× speedup and higher task success.VFlash employs Flash Primitive DecodingFlash Primitive DecodingThe technique of jointly decoding all coordinates of a visual primitive in one model forward pass. to jointly predict visual primitivevisual primitiveA structured visual element such as a point, bounding box, or trajectory used to represent spatial goals. coordinates in a single forward pass, yielding ~25× faster generation and higher task success.Curator's notesWritten by AI from the paperEvery number here appears in the abstractReport a mistake
What's new
It replaces token‑by‑token autoregressiveautoregressiveA generation method that produces output tokens sequentially, each conditioned on previous tokens. decoding with a single‑pass drafter that predicts whole primitives jointly.
Key result
A world action model using VFlash plans reaches 70.2% average success across four MolmoSpaces tasks, versus 62.6% baseline.
Why it matters
Builders get much quicker and more precise visual goal predictions, enabling responsive robotics and autonomous systems.
Key numbers
From the paperCopied from the abstract, not written by AI.
Abstract
From the paperThe authors' own words.
Models for physical AI tasks, such as robotics and autonomous driving, need not reason solely in language. They can use visual plans--points, bounding boxes, and trajectories--to represent spatial goals and anticipate motion. However, autoregressive vision-language models generate these plans one coordinate token at a time, making richer predictions increasingly costly. We introduce VFlash, a method for fast, accurate visual planning through Flash Primitive DecodingFlash Primitive DecodingThe technique of jointly decoding all coordinates of a visual primitive in one model forward pass.. VFlash exploits the known structure of visual primitives to predict their coordinates jointly, using a lightweight drafter that generates an entire primitive in one forward pass. The layout adapts to the primitive's type, dimensionality, and length, supporting compact boxes and long, variable-length trajectories. Crucially, spatial accuracy need not require reproducing an autoregressive model's exact token sequence. VFlash therefore trains on annotated primitives and directly accepts structurally valid predictions without target-token agreement. Across the evaluated 2D and 3D robot trajectory tasks, VFlash consistently reduces prediction errors relative to autoregressive decoding. Preliminary single-rollout measurements indicate approximately 25x faster generation for 3D point motion forecasting. A world action model conditioned on VFlash plans achieves 70.2% average success across four MolmoSpaces task suites, compared with 62.6% for the baseline.
From the full paper
From the paperCopied from the paper itself by our code.
Reading the paper on arXiv…
What you could build
Written by AIIntermediateRequires training or fine‑tuning vision models and integrating the joint decoder
- Generate real‑time 3D trajectory forecasts for drone navigation
- Create an interactive robot arm planner that outputs bounding‑box goals instantly
- Build a visual‑language assistant that suggests motion plans for self‑driving cars
Ask this paper
/ Ask a questionAnswers come from the full paper and name the section
/ Quiz yourself
How does VFlash generate visual primitives compared to autoregressive models?
VFlash uses a lightweight drafter to predict all coordinates of a primitive jointly in one forward pass, instead of generating tokens sequentially.
/ Your notes
Select text in the abstract to highlight it or add a note. They stay in this browser.