US11429894B2
patent
Example aspects of the present disclosure are directed to systems and methods for learning classification models which satisfy constraints such as , for example, constraints that can be expressed as a predicted positive rate or negative rate on a subset of the training dataset . In particular, through the use of quantile estimators , the systems and methods of the present disclosure can transform a constrained optimi zation problem into an unconstrained optimization problem that is solved more efficiently and generally than the con strained optimization problem . As one example, the uncon strained optimization problem can include optimizing an objective function where aa decision threshold of the classi fication model is expressed as an estimator of a quantile function on the classification scores of the machine - learned
18 claims. 96,609 characters of text.