Compositional Foundation Models News Today

A multimodal system uses models trained on language, vision and action data helps robots develop and execute plans for manufacturing tasks.

Your daily to-do list is likely pretty straightforward: wash the dishes, buy groceries, and other minutiae. It’s unlikely you wrote out “pick up the first dirty dish,” or “wash that plate with a sponge,” because each of these miniature steps within t

Construction industry news, trends and jobs for building professionals who want mobile-friendly content.

Your daily to-do list is probably pretty simple: washing dishes, grocery shopping, and other minutiae. It's unlikely you wrote "pick up the first dirty plate"

MIT's HiP system helps robots complete long-horizon goals using three foundation models: a large language model, a video diffusion model, and an egocentric action model. Iterative refinement improves the plan at each step for household and manufacturing tasks.