Regularization

A class of techniques applied to ill-posed or unstable inverse problems and statistical estimation that impose additional constraints or penalties to stabilize solutions and control complexity.

A family of techniques that stabilize ill-posed or ill-conditioned problems by adding information, constraints, or smoothing terms (regularizers) to select a well-behaved solution; commonly used in inverse problems, optimization, and machine learning.

A family of procedures that modify a mathematical object or problem to remove singularities or ill-posedness, typically by adding smoothing, penalization, or constraints so that the modified problem is stable and better behaved.