Spatial intelligence lets an agent perceive and act in the physical world, and many of its judgments must be made at interaction rates: which spatial relation holds, how far an object is, or where to place it. Vision-language models (VLMs) produce these answers by free-form text generation, paying for autoregressive de...
Luo-Zheng Qin, Yu-Han Zhu, Chang-Hai Zhou et al.· 0 citations
CoBind is introduced, a training-free framework for stage-aware compositional binding that parses a prompt into a composition graph of entities, attributes, and relations and adapts the guidance strength according to the current satisfaction of each constraint, reducing unnecessary latent updates.
Kaijie Chen, Ethan Caldwell, M. Vossen et al.· arXiv.org· 0 citations
This work presents LongStraw, an objective-aware, architecture-aware system for resident-state virtualization, response replay, and distributed-gradient execution that bounds the live training graph by the response suffix while reusing the expensive prompt computation across the complete GRPO group.
