Mini-Batch Gradient Descent

Computing gradients on small batches of data, balancing SGD’s noise with full-batch GD’s stability.

This concept is essential for understanding training & optimization and forms a key part of modern AI systems.