A generalization of scaling that modifies the hessian so as to improve convergence properties In the case of quadratic programming, this typically means multiplying the quadratic form matrix, Q, by a preconditioner, P, so that the condition number of PQ is as close to 1 as possible There are many variations, including splitting a matrix, Q = P - S, to achieve a better conditioned matrix in the sense of its eigenvalue (and eigenspace) structure that governs convergence properties of algorithms like steepest ascent and conjugate gradient
{i} advance training, adapting or habituating ahead of time, developing conditioned responses in advance
an assertion that must hold prior to the successful execution of the associated function (e g , use case path, class method) See also invariant and postcondition
If one thing is a precondition for another, it must happen or be done before the second thing can happen or exist. They made multi-party democracy a precondition for giving aid. something that must happen or exist before something else can happen precondition of/for