Pooling Layer
A layer in neural networks that reduces spatial dimensions by aggregating values in local regions, decreasing computation and providing translation invariance.
Types
Max pooling: Takes the maximum value in each region. Average pooling: Takes the mean. Global average pooling: Reduces entire spatial dimension to a single value per channel. Adaptive pooling: Outputs a specified size regardless of input.
Modern Trends
Max pooling was standard in older CNNs (AlexNet, VGG). Modern architectures often use strided convolutions instead. Vision Transformers don't use pooling at all, replacing it with class tokens and attention.