Zhejiang Univ's ProVisE: Evaluating Spatial Cognition via Generative Pixels
机器之心 · wechat · 2026-08-09
The OmniAI team at Zhejiang University introduced ProVisE, a framework that shifts spatial cognition evaluation away from text-based coordinates by allowing generative models to "draw" answers directly in pixel space.
ProVisE uses visual protocols to guide models in marking targets, generating depth maps, or drawing trajectories, which are then parsed into structured predictions. The team also released SpatialGen-Bench, covering 14 sub-tasks to comprehensively assess spatial cognition.
Experiments reveal that generative models exhibit stronger "spatial intuition" for direct visual perception, complementing text-based VLMs. However, text VLMs still hold an advantage in tasks requiring abstract reasoning.
Related event: Zhejiang Univ Introduces ProVisE for AI Spatial Reasoning(2 posts)→
More from Research
- Simplex Diffusion Models paper talk by EPFL/DeepMind's Justin Deschenaux on Oct 19 — ArnaudDoucet1 · 2026-10-16
- Meituan LongCat team's Data World synthesizes data environments to train data agents — meituan-longcat · 2026-10-16
- FDE-Gym benchmark shows huge gap on real AI delivery tasks: top model scores 0.563 — DeepLethe · 2026-10-16
- ForcingBooster halves computation of autoregressive video diffusion with patch-wise denoising — VAST-AI · 2026-10-16
- Baidu open-sources FlashMask: up to 1.55x faster than FlashAttention-4 for long-context training — BaiduResearch · 2026-10-16
- SemEval-2027 Task 9 opens: multimodal news framing analysis across 4 languages — preslav_nakov · 2026-10-16