<ul><li>Prior work primarily formulates Context-Aware Human Activity Recognition (CA-HAR) as a multi-label classification problem.</li><li>Existing CA-HAR methods struggle to capture the semantic relationships between activity labels, limiting their accuracy.</li><li>To address this limitation, researchers propose SEAL, which leverages Language Models (LMs) to encode CA-HAR activity labels.</li><li>SEAL uses LMs to generate vector embeddings that preserve rich semantic information from natural language, improving CA-HAR accuracy.</li></ul>

Semantically Encoding Activity Labels for Context-Aware Human Activity Recognition

Discover more