We show how LLM prompts can be algorithmically generated for diverse human-like coordination and communication. Super excited about its implications for scenario generation and training better embodied assistants in the future!
One of the biggest challenges in designing agents/robots that collaborate effectively with humans is the lack of diverse human data—large scale collection is hard!
Introducing PLAN-QD: a framework that algorithmically generates diverse, human-like agents.
👇 Learn more! [1/9]
