Problem
Research questions and friction points this paper is trying to address.
zero-shot learning
Wi-Fi-based human activity recognition
unseen activity recognition
closed-set assumption
labeled Wi-Fi samples
Zero-Fi: Zero-Shot Wi-Fi-Based Human Activity Recognition via Contrastive Signal-Language Alignment
🤖 AI Summary
This work addresses the limited generalizability of conventional Wi-Fi-based human activity recognition methods, which rely on predefined and fully annotated activity categories and struggle to recognize unseen activities. To overcome this limitation, the authors propose Zero-Fi, a novel framework that introduces cross-modal alignment between Wi-Fi signals and natural language for the first time in wireless sensing. By leveraging contrastive learning, Zero-Fi maps Wi-Fi signal features and textual activity descriptions into a shared semantic embedding space, enabling zero-shot activity recognition. The approach requires neither labeled data nor model fine-tuning for new activity categories. Extensive experiments on large-scale public datasets demonstrate that Zero-Fi effectively recognizes previously unseen activities within held-out categories, substantially expanding the practical applicability of Wi-Fi sensing.
📝 Abstract
Wi-Fi-based human activity recognition has advanced substantially, but most existing methods assume a closed set of activities and require labeled Wi-Fi samples for every target class, limiting their ability to recognize unseen activities. We present Zero-Fi, a contrastive signal-language alignment framework for zero-shot Wi-Fi-based human activity recognition. Zero-Fi learns unified representations from complementary Wi-Fi signal features and aligns them with the semantic representations of natural-language activity descriptions in a shared embedding space. This cross-modal alignment enables Zero-Fi to recognize new activity classes without requiring labeled Wi-Fi samples or model adaptation for those classes. Experiments on large-scale public benchmark datasets demonstrate effective zero-shot recognition of held-out activity classes, highlighting the potential of signal-language alignment to extend Wi-Fi sensing beyond predefined activity classes.
Innovation
Methods, ideas, or system contributions that make the work stand out.
zero-shot learning
Wi-Fi sensing
contrastive alignment
signal-language representation
human activity recognition
🔎 Similar Papers
No similar papers found.
💼 Related Jobs
