Batch Normalization

Quick Definition

Normalizing layer inputs to stabilize and speed up training

A technique normalizing layer inputs during training to stabilize and accelerate neural network convergence.

faster training, reduced covariate shift, higher learning rates

deep-learning neural-network gradient-descent

Distance between decision boundary and nearest data points

Identifying the most relevant features for model training

Stochastic model where events depend only on previous state

Fast gradient boosting framework using histogram splitting

Ensemble of decision trees that votes on predictions

Finding optimal hyperplane to separate classes with max margin