Search papers, labs, and topics across Lattice.
This paper introduces LEEVLA, a novel architecture for vision-language-action (VLA) models that enhances performance in complex dynamic scenarios by emphasizing task-critical evidence. The method employs drift-guided dynamic prioritization (DGDP) to direct the model's attention to salient regions and structured feature flow generation (SFFG) to ensure consistent evolution of latent representations. Experimental results demonstrate that LEEVLA significantly outperforms existing approaches, highlighting the importance of explicit task-evidence guidance and structured reasoning in VLA systems.
LEEVLA reveals that effectively guiding attention to task-critical evidence can dramatically enhance performance in vision-language-action tasks.
Vision-language-action (VLA) models aim to map multimodal inputs to robot actions. However, most existing approaches struggle to cover complex dynamic scenarios due to treating all visual tokens uniformly and reasoning with human-selected factors, which lack mechanisms to emphasize task-critical evidence and ignore underlying factors. To address this issue, we propose LEEVLA, a VLA architecture for seeing what matters in Latent Environment Evolution that explicitly guides the model toward informative regions while preserving the structured evolution of latent world representations. To identify salient and instruction-relevant regions, we introduce drift-guided dynamic prioritization (DGDP), which combines dynamic position prioritization (DPP) with semantic drift guidance (SDG) to guide the VLA agent where to attend during training. On top of this, we introduce structured feature flow generation (SFFG), which models how these prioritized features should evolve in latent space via prototype-to-periphery (P2P) prediction, and a mutual-neighborhood contrastive (MC) loss to maintain topological consistency among neighborhoods. Together, DGDP and SFFG form a task-aware"where-how"training framework. Extensive experiments on VLA benchmarks show that LEEVLA consistently outperforms prior methods, confirming that explicit task-evidence guidance and structured latent reasoning are both crucial for scalable VLA. Our code is available at https://github.com/LyuQi127/LEEVLA.
