Meta-learning, often referred to as "learning to learn," is a machine learning paradigm in which a model is trained across a distribution of tasks so that it can rapidly adapt to new, unseen tasks with minimal labeled data or few gradient update steps. In the context of WiFi CSI-based sensing, meta-learning is particularly valuable because real-world deployment scenarios frequently involve domain shifts — such as changes in environment, user, or device — where collecting large amounts of labeled data for each new setting is impractical. Key variants relevant to this field include model-agnostic meta-learning (MAML), which optimizes model initialization for fast fine-tuning, and metric-based approaches such as prototypical networks, both of which support efficient, generalizable sensing under the data-scarce and energy-constrained conditions central to green wireless sensing research.
Source Papers
- DASECount: Domain-Agnostic Sample-Efficient Wireless Indoor Crowd Counting via Few-Shot Learning ↗ — DASECount: Domain-Agnostic Sample-Efficient Wireless Indoor
- Deep Learning-Enhanced Human Sensing with Channel State Information: A Survey ↗ — Deep Learning-Enhanced Human Sensing with Channel State Info