SAGE: Smart home Agent with Grounded Execution

2311.00772v1

This article introduces SAGE (Smart home Agent with Grounded Execution), a framework designed to maximize the flexibility of smart home assistants by replacing manually-defined inference logic with an LLM-powered autonomous agent system. SAGE integrates information about user preferences, device states, and external factors (such as weather and TV schedules) through the orchestration of a collection of tools. SAGE’s capabilities include learning user preferences from natural-language utterances, interacting with devices by reading their API documentation, writing code to continuously monitor devices, and understanding natural device references. To evaluate SAGE, we develop a benchmark of 43 highly challenging smart home tasks, where SAGE successfully achieves 23 tasks, significantly outperforming existing LLM-enabled baselines (5/43).

None