System

A system comprising:

• a communication interface configured to communicate with a client terminal via a packet-switched network conforming to at least one of a Fifth Generation Mobile Communication System, Wi-Fi, or Bluetooth communication standard;
• a storage device storing a data generation model comprising a transformer-based neural network obtained by deep learning, and an emotion identification model;
• and
• circuitry configured to:
• receive, from the client terminal via the communication interface, input data comprising a natural-language character string representing a user request;
• perform normalization preprocessing on the input data, the normalization preprocessing comprising character-type unification, encoding normalization, removal of unnecessary blank characters and control symbols, and conversion of full-width characters into half-width characters;
• generate a task-specific prompt by embedding preprocessed data obtained from the normalization preprocessing into an instruction template, the instruction template specifying a structured output schema that defines a plurality of extraction fields as key-value pairs;
• transmit the task-specific prompt to the data generation model and receive response data comprising extracted values formatted as a serialized data structure;
• validate the extracted values by verifying a format and a value range for each extraction field, and iteratively refine the task-specific prompt when a missing or invalid field is detected;
• estimate an emotion of a user by applying the emotion identification model to sensor data received from the client terminal;
• and
• transmit, to the client terminal via the communication interface, output data generated based on the validated extracted values, a presentation format of the output data being adjusted based on the estimated emotion.

A method performed by a system comprising a communication interface and circuitry, the method comprising:

• receiving, from a client terminal via the communication interface and a packet-switched network, input data comprising a natural-language character string;
• performing normalization preprocessing on the input data, the normalization preprocessing comprising at least character-type unification and encoding normalization, to generate preprocessed data;
• generating a task-specific prompt by embedding the preprocessed data into an instruction template that specifies a structured output schema defining a plurality of extraction fields;
• transmitting the task-specific prompt to a data generation model stored in a storage device of the system, and receiving, from the data generation model, response data comprising extracted values corresponding to the plurality of extraction fields;
• validating the extracted values by verifying a format and a value range for each of the plurality of extraction fields, and, when a missing or invalid field is detected, generating a modified prompt and retransmitting the modified prompt to the data generation model;
• and
• transmitting, to the client terminal via the communication interface, output data generated based on the validated extracted values.