Stochastic Gradient Descent (SGD) is an optimization algorithm used to minimize a function by iteratively updating its parameters. Unlike traditional gradient descent, which uses the entire dataset to compute gradients, SGD randomly selects a single data point or a small batch of data points for each update. This makes the process faster and allows the algorithm to start improving the model immediately. SGD is particularly useful in training machine learning models, such as neural networks, where large datasets can make traditional methods computationally expensive. By using random samples, SGD introduces variability, which can help escape local minima and lead to better overall solutions.