Penalizing model complexity to fight overfitting — L1/L2 weight penalties and dropout.
Reference Q&A for this concept is coming soon.
