Using a neural network to optimize processing of user requests

A system for using a neural network to select customer requests for assignment, the system comprising:

• at least one server computer comprising at least one processor and at least one memory, the at least one server computer configured to:
• obtain a first selection model, wherein the first selection model comprises the neural network and processes feature vectors corresponding to customer requests and generates scores for selection of customer requests for assignment;
• use the first selection model during a first time period to select customer requests, wherein, during the first time period, a plurality of selection decisions are made, and wherein a first selection decision of the plurality of selection decisions comprises:
• determining that a first customer service representative is available for assignment,
• obtaining information about a plurality of customer requests awaiting assignment,
• computing a probability distribution for the plurality of customer requests using the first selection model, wherein computing the probability distribution comprises computing a selection score for a first customer request by (i) creating a first feature vector using information about the first customer request and (ii) processing the first feature vector using the first selection model,
• sampling the probability distribution to select the first customer request from the plurality of customer requests for assignment to the first customer service representative, and
• assigning the first customer service representative the selected first customer request;
• compute a reward score wherein the reward score corresponds to a performance of the first selection model during the first time period, wherein the reward score is computed by comparing the performance of the first selection model during the first time period to the performance of a third selection model during a third time period, wherein the third time period is prior to the first time period;
• compute a second selection model by modifying parameters of the first selection model using the reward score and the plurality of selection decisions;
• and
• use the second selection model during a second time period to select a second customer request for assignment to a second customer service representative.

One or more non-transitory, computer-readable media comprising computer executable instructions that, when executed, cause at least one processor to perform actions comprising:

• obtaining a first selection model, wherein the first selection model processes feature vectors corresponding to customer requests and generates scores for selection of customer requests for assignment;
• using the first selection model during a first time period to select customer requests, wherein, during the first time period, a plurality of selection decisions are made, and wherein a first selection decision of the plurality of selection decisions comprises:
• determining that a first customer service representative is available for assignment,
• obtaining information about a plurality of customer requests awaiting assignment,
• computing a probability distribution for the plurality of customer requests using the first selection model, wherein computing the probability distribution comprises computing a selection score for a first customer request by (i) creating a first feature vector using information about the first customer request and (ii) processing the first feature vector using the first selection model,
• sampling the probability distribution to select the first customer request from the plurality of customer requests for assignment to the first customer service representative, and
• assigning the first customer service representative to the selected first customer request;
• computing a reward score wherein the reward score indicates a performance of the first selection model during the first time period, wherein the reward score is computed by comparing the performance of the first selection model during the first time period to the performance of a third selection model during a third time period, wherein the third time period is prior to the first time period;
• computing a second selection model by modifying parameters of the first selection model using the reward score and the plurality of selection decisions;
• and
• using the second selection model during a second time period to select a second customer request for assignment to a second customer service representative.