Tuning external invocations utilizing weight-based parameter resampling

A computer system, comprising:

• a memory configured to store computer-executable instructions;
• and
• one or more processors in communication with the memory and configured to access the memory and execute the computer-executable instructions to at least:
• determining a plurality of weights respectively for a plurality of parameter values, wherein a weight of each particular parameter value of the plurality of parameter values is associated with a probability of selecting a sample having the particular parameter value and wherein each respective value of the plurality of parameter values is unique;
• randomly selecting, based at least in part on the plurality of initial weights, a first parameter value of the plurality of parameter values, the first parameter value to be used for selecting a first sample;
• selecting the first sample of a plurality of samples corresponding to a first batch of data records for processing via a first external invocation;
• invoking the processing of the first sample via the first external invocation;
• determining feedback data associated with a level of performance of the first external invocation;
• adjusting the initial weights of the plurality of parameter values based at least in part on the feedback data;
• randomly selecting, based at least in part on the adjusted plurality of weights, a second parameter value of the plurality of parameter values, the second parameter value to be used for selecting a second sample;
• selecting the second sample of the plurality of samples corresponding to a second batch of data records according to the second parameter value, for processing via a second external invocation;
• and
• invoking the processing of the second sample via the second external invocation;
• wherein the plurality of parameter values remains fixed between external invocations that are executed to process the plurality of data records.

A method, comprising:

• determining, by a computing system, a plurality of weights respectively for a plurality of parameter values, wherein a weight of each particular parameter value of the plurality of parameter values is associated with a probability of selecting a sample having the particular parameter value and wherein each respective value of the plurality of parameter values is unique;
• randomly selecting, by the computing system and based at least in part on the plurality of initial weights, a first parameter value of the plurality of parameter values, the first parameter value to be used for selecting a first sample;
• selecting, by the computing system, the first sample of a plurality of samples corresponding to a first batch of data records for processing via a first external invocation;
• invoking the processing of the first sample via the first external invocation;
• determining, by the computing system, feedback data associated with a level of performance of the first external invocation;
• adjusting, by the computing system, the initial weights of the plurality of parameter values based at least in part on the feedback data;
• randomly selecting, by the computer system and based at least in part on the adjusted plurality of weights, a second parameter value of the plurality of parameter values, the second parameter value to be used for selecting a second sample;
• selecting, by the computer system, the second sample of the plurality of samples corresponding to a second batch of data records according to the second parameter value, for processing via a second external invocation;
• and
• invoking the processing of the second sample via the second external invocation;
• wherein the plurality of parameter values remains fixed between external invocations that are executed to process the plurality of data records.