System

A system comprising:

• circuitry configured to:
• receive, via a packet-switched network, input data records each including a character string field and a temporal attribute field;
• transmit the character string field to an external information processing apparatus having a natural language processing function and acquire analysis result data including topic information, sentiment information, keyword information, and date-time expression information;
• construct a structured prompt based on the character string field, the temporal attribute field, and at least a portion of the analysis result data, the structured prompt including an instruction portion that specifies extraction of a priority score value, an action classification value, a temporal constraint datum, and a condensed representation datum;
• transmit the structured prompt to a generative neural network model via an application programming interface and acquire inference result data conforming to the instruction portion;
• compute a feature vector from the input data records and the analysis result data, the feature vector including a sentiment index, a length index, a classification index, an urgency keyword count, and an interrogative marker count;
• determine, using a trained classification model and the feature vector, the action classification value as one of a reply class, a task class, or a no-action class;
• accumulate behavior history data of a user and update a user-specific machine learning model based on the behavior history data to generate a personalized priority score value;
• and
• generate output signal data including the personalized priority score value, the action classification value, the temporal constraint datum, and the condensed representation datum, and transmit the output signal data to a terminal device via the packet-switched network.

A method performed by circuitry, the method comprising:

• receiving, via a packet-switched network, input data records each including at least a character string field and a temporal attribute field;
• constructing a structured prompt based on at least the character string field and the temporal attribute field, the structured prompt including an instruction portion that specifies a plurality of extraction tasks;
• transmitting the structured prompt to a generative neural network model via an application programming interface;
• acquiring, from the generative neural network model, inference result data including a priority score value, an action classification value, a temporal constraint datum, and a condensed representation datum;
• computing a feature vector from the input data records and the inference result data, the feature vector including at least a sentiment index, a length index, and a classification index;
• and
• generating output signal data based on the priority score value and the condensed representation datum, and transmitting the output signal data to a terminal device via the packet-switched network.