1 paper
Di Wu, Jiaxin Fan, Junzhe Zang +4
Embodied planning requires agents to make coherent multi-step decisions based on dynamic visual observations and natural language goals. While recent vision-language models (VLMs)…