Misplaced glasses on the kitchen counter or a phone left somewhere in the living room could soon be found by a robot.

There have been many advances in vision-language models (VLM) that can match natural language queries to objects in a visual scene. And researchers are experimenting with how these models can be ...