Softmax is a mathematical activation function applied at the output layer of a neural network that converts a vector of raw scores (logits) into a probability distribution over multiple classes, ensuring all outputs sum to one. In Wi-Fi CSI sensing tasks such as occupancy detection and passenger counting, softmax enables the model to produce interpretable confidence scores for each class, making it well-suited for multi-class classification problems where mutually exclusive outcomes must be ranked. The standard softmax is the most common variant used in these contexts, though temperature-scaled softmax is sometimes employed to calibrate output confidence, and log-softmax is often paired with negative log-likelihood loss for numerical stability during training.
Source Papers
- Channel State Information (CSI) Amplitude Coloring Scheme for Enhancing Accuracy of an Indoor Occupancy Detection System Using Wi-Fi Sensing ↗ — Channel State Information (CSI) Amplitude Coloring Scheme fo
- RSSI-Assisted CSI-Based Passenger Counting with Multiple Wi-Fi Receivers ↗ — RSSI-Assisted CSI-Based Passenger Counting with Multiple Wi-