MuRAL: A Multi-Resident Ambient Sensor Dataset Annotated with Natural Language for Activities of Daily Living

Researchers have released MuRAL, a new dataset containing over 21 hours of multi-resident smart home sensor data with natural language annotations for training AI models. The dataset aims to improve Large Language Models' ability to understand human activities in complex smart home environments, though current LLMs still struggle with key tasks like resident identification and activity prediction.

• →MuRAL dataset provides 21+ hours of multi-user smart home sensor data with detailed natural language descriptions and activity labels.
• →Current widely-used datasets like CASAS, ARAS, and MARBLE lack natural language context needed for modern LLM training.
• →Benchmarking results show existing LLMs face significant challenges in maintaining accurate resident assignment over long sequences.
• →The dataset addresses gaps in fine-grained annotation for realistic multi-resident smart environment scenarios.
• →MuRAL is publicly available and designed to advance human activity recognition research using ambient sensors.