Velox: Spatial Memory Makes World Models Efficient
Read the original on arXiv Computer Vision →Velox is a training‑free attention router that improves the efficiency of video world models by selecting region‑level context through frustum slab correspondences derived from streaming geometry. It represents past blocks as camera frustum slabs with depth intervals and uses a recency‑priority multiwarp to obtain current‑block depth, soft‑modulating importance for well‑covered regions while preserving model‑derived routing for uncovered areas. The fixed token‑block granularity aligns with FlashAttention‑style tiled execution, reducing the attention term that grows with rollout length and achieving significant speedups—4.36× acceleration in attention and 2.43× in denoising at chunk 48—while maintaining performance close to dense models on the WBench navigation split.
Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Computer Vision.