6 papers - avg viability 6.3
DepthCache is a training-free framework that optimizes visual token merging for faster robotic manipulation without degrading performance.
AR-VLA is a context-aware autoregressive action generator for robotic manipulation tasks that enhances action trajectory smoothness and task success rates.
Fast-ThinkAct optimizes embodied AI for faster and smarter action execution in visual-language tasks.
ReMem-VLA enhances robot control by integrating advanced memory mechanisms into vision-language-action models.
KineVLA enhances robotic manipulation through a novel vision-language-action framework that integrates detailed kinematic attributes.
DeepVision-VLA enhances visual representations in Vision-Language-Action models for improved robotic manipulation.