Performance metric based stopping criteria for iterative algorithms

In a computer environment, a system of identifying a stopping criterion for an iterative algorithm based on more than a cost function, the system comprising:

• one or more modules implemented at least partially by a processing system and embodied within a non-transitory computer-readable storage medium, the computer-readable storage medium storing instruction that, responsive to execution by the processing system, causes operations to be performed comprising:
• processing a training dataset using the iterative algorithm having the cost function, the processing including for a plurality of iterations of the iterative algorithm, computing:
• a cost for the iterative algorithm using the cost function;
• and
• a value for each of a plurality of performance metrics that are separate and distinct from the cost function and that are usable to infer accuracy of the iterative algorithm for a respective said iteration;
• responsive to the processing, identifying the stopping criterion based at least in part on the computed values for the plurality of performance metrics that are not the cost function;
• and
• outputting the stopping criterion to configure the iterative algorithm and use the stopping criterion for subsequent processing of data by the iterative algorithm.