HELP: Hierarchical Embodied Language Planner for Household Tasks

This paper addresses the challenge of enabling embodied agents to perform complex household tasks by leveraging the power of Large Language Models (LLMs). The key contribution is the development of a hierarchical planning architecture (HELP) that decomposes complex tasks into subtasks, allowing LLMs to handle linguistic ambiguity and environmental interactions effectively. The focus on using open-source LLMs with fewer parameters is significant for practical deployment and accessibility.