Batch normalization
Also: batch norm, BatchNorm, BN, batch normalisation
A layer that standardizes each activation using the mini-batch mean and variance, then applies a learned scale γ and shift β. It stabilizes and speeds up training, allows higher learning rates, and has a mild regularizing effect; at inference it uses running averages instead of batch statistics.
y = γ · (x − μ_B) / √(σ²_B + ε) + β
Example
A feature with batch values (1, 2, 3) has mean 2 and variance 2/3, so it is normalized to about (−1.22, 0, 1.22) before γ and β rescale it.