# US 2026/0289081 A1 · System

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2026/0289081 A1
- Document ID: US20260289081A1
- Status: Published application
- Kind: A1
- Application number: 19/562,878
- Filed: 2026-03-11
- Published: 2026-09-24
- Earliest claimed: 2025-03-19
- Assignees: SoftBank Group Corp.
- Applicants: SoftBank Group Corp.
- Inventors: Yuko WATANABE
- Web page: https://patentgrep.com/patent/US20260289081A1

## CPC

- G06F 40/166: Editing, e.g. inserting or deleting
- G06F 40/103: Formatting, i.e. changing of presentation of documents (automatic justification G06F40/189; automatic line break hyphenation G06F40/191)
- G06F 40/226: Validation
- G06N 3/045: Combinations of networks
- G06N 3/0475: Generative networks
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

A system includes a processor that is configured to receive a request input by a user and convert the request into a bullet-point format by using a natural language processing algorithm, create, based on the request converted into the bullet-point format, a prompt for instructing a generative artificial intelligence model to generate a natural language sentence, and present the generated natural language sentence to the user via a user interface and provide an editing tool that allows the user to review the sentence and modify the sentence as needed.

## Claims

### Claim 1

A system comprising:
circuitry configured to:
receive, from a terminal device via a communication interface coupled to a packet-switched network, input data representing a plurality of items, and convert the input data into structured data in a list format by executing a natural language processing algorithm;
generate, based on the structured data, a prompt sentence for instructing a generative neural network model to generate a natural language text, the prompt sentence including at least the structured data and an instruction portion specifying generation conditions;
transmit the prompt sentence to the generative neural network model, acquire a generated natural language text from the generative neural network model, and perform post-processing on the generated natural language text; and
transmit response data including the post-processed natural language text to the terminal device via the communication interface and the packet-switched network, the response data causing the terminal device to present the post-processed natural language text in an editable region of a user interface.

### Claim 2 (depends on 1)

The system according to claim 1, wherein the circuitry is further configured to perform format validation and content validation on the structured data before generating the prompt sentence, the format validation including verifying that a number of items in the structured data is within a predetermined range and that each item does not exceed a predefined character limit, and the content validation including scanning the items for prohibited character sequences using pattern matching.

### Claim 3 (depends on 2)

The system according to claim 2, wherein the circuitry is further configured to discard or flag items that violate the format validation or the content validation, and to generate an internally validated representation of the structured data for use in generating the prompt sentence.

### Claim 4 (depends on 1)

The system according to claim 1, wherein the prompt sentence further includes a fixed instruction sentence retrieved from a template stored in a storage device, style conditions specifying at least one of a language, a speaker viewpoint, a politeness level, and a document length, and output format conditions specifying a structure of the generated natural language text.

### Claim 5 (depends on 4)

The system according to claim 4, wherein the circuitry is further configured to set generation parameters for the generative neural network model, the generation parameters including at least one of a temperature value, a maximum token number, and a sampling strategy, and to generate model input data by combining the prompt sentence with the generation parameters.

### Claim 6 (depends on 5)

The system according to claim 5, wherein the generative neural network model comprises a transformer-based neural network including an embedding layer, a plurality of self-attention layers, and an output projection layer, and wherein the generative neural network model is trained using a cross-entropy loss function and a gradient descent optimization algorithm.

### Claim 7 (depends on 6)

The system according to claim 6, wherein training of the generative neural network model includes data augmentation techniques comprising at least one of token masking and random span corruption.

### Claim 8 (depends on 1)

The system according to claim 1, wherein the post-processing includes at least one of normalization of line breaks, normalization of character encoding, enforcement of a consistent sentence-ending punctuation pattern, and truncation of the generated natural language text when the generated natural language text exceeds a predetermined size.

### Claim 9 (depends on 8)

The system according to claim 8, wherein the response data further includes the plurality of items of the structured data and the prompt sentence, and the circuitry is further configured to include metadata comprising at least one of a model identifier, a generation timestamp, and token usage statistics in the response data.

### Claim 10 (depends on 1)

The system according to claim 1, wherein the circuitry is further configured to receive, from the terminal device via the communication interface, editing information indicating modifications made by an operator to the post-processed natural language text in the editable region, and to determine, based on the editing information, whether the post-processed natural language text was modified.

### Claim 11 (depends on 10)

The system according to claim 10, wherein the circuitry is further configured to, when the post-processed natural language text was not modified, automatically treat the post-processed natural language text as finalized data, and to, when the post-processed natural language text was modified, use an edited version as the finalized data.

### Claim 12 (depends on 11)

The system according to claim 11, wherein the circuitry is further configured to convert the finalized data into transmission data and transmit the transmission data to another information processing apparatus via the communication interface and the packet-switched network.

### Claim 13 (depends on 12)

The system according to claim 12, wherein the circuitry is further configured to perform, on the finalized data, a classification operation using at least one of a classification model and the generative neural network model to determine a category label, and to determine a processing destination or a processing priority for the transmission data based on the category label.

### Claim 14 (depends on 10)

The system according to claim 10, wherein the circuitry is further configured to acquire a history of editing operations performed on the post-processed natural language text and accumulate the history as learning information used to update at least one of a configuration of the prompt sentence and a generation condition of the generative neural network model.

### Claim 15 (depends on 1)

The system according to claim 1, wherein the circuitry is further configured to estimate an emotion of an operator by applying an emotion identification model to sensor data received from the terminal device via the communication interface, and to adjust at least one of the prompt sentence and the generation conditions based on the estimated emotion.

### Claim 16 (depends on 15)

The system according to claim 15, wherein the emotion identification model comprises a neural network trained on training data sets that each combine a user input with an emotion value mapped on an emotion map, and wherein the emotion identification model classifies the emotion into at least one of a plurality of emotion categories.

### Claim 17 (depends on 15)

The system according to claim 15, wherein the sensor data includes at least one of audio data representing speech of the operator, image data representing a facial expression of the operator, and text data representing textual input of the operator, and wherein the emotion identification model performs multi-modal emotion analysis on the sensor data.

### Claim 18

A system comprising:
circuitry configured to:
receive, from a terminal device via a communication interface coupled to a packet-switched network, input data representing a plurality of requirement items input by an operator, and convert the input data into structured data in a bullet-point format by executing a natural language processing algorithm including at least one of tokenization, parsing, and entity extraction;
perform format validation including verifying that a number of the requirement items is within a predetermined range and content validation including scanning for prohibited character sequences, and generate a validated representation of the structured data;
generate, based on the validated representation, a prompt sentence for instructing a generative neural network model to generate a natural language inquiry sentence, the prompt sentence including a fixed instruction sentence, style conditions specifying at least one of a language, a speaker viewpoint, and a politeness level, output format conditions, and the requirement items embedded in a list format;
set generation parameters including at least one of a temperature value, a maximum token number, and a sampling strategy, transmit model input data including the prompt sentence and the generation parameters to the generative neural network model, acquire a generated natural language inquiry sentence, and perform post-processing including normalization of line breaks, character encoding, and sentence-ending expressions;
transmit response data including the post-processed natural language inquiry sentence, the requirement items, and the prompt sentence to the terminal device via the communication interface and the packet-switched network, the response data causing the terminal device to present the post-processed natural language inquiry sentence in an editable input area;
receive editing information from the terminal device indicating whether the operator modified the post-processed natural language inquiry sentence, determine a finalized natural language text based on the editing information, and convert the finalized natural language text into transmission data; and
transmit the transmission data as inquiry information to another information processing apparatus via the communication interface and the packet-switched network.

### Claim 19 (depends on 18)

The system according to claim 18, wherein the generative neural network model comprises a transformer-based neural network including an embedding layer that maps discrete tokens to vector representations, a sequence of self-attention and feed-forward layers, and an output projection layer that converts hidden state vectors into probability distributions over a token vocabulary, and wherein the generative neural network model is trained by minimizing a cross-entropy loss between predicted tokens and reference tokens using gradient descent, with data augmentation comprising at least one of token masking and random span corruption.

### Claim 20

A method performed by a system comprising circuitry, the method comprising:
receiving, from a terminal device via a communication interface coupled to a packet-switched network, input data representing a plurality of items, and converting the input data into structured data in a list format by executing a natural language processing algorithm;
generating, based on the structured data, a prompt sentence for instructing a generative neural network model to generate a natural language text, the prompt sentence including at least the structured data and an instruction portion specifying generation conditions;
transmitting the prompt sentence to the generative neural network model, acquiring a generated natural language text from the generative neural network model, and performing post-processing on the generated natural language text; and
transmitting response data including the post-processed natural language text to the terminal device via the communication interface and the packet-switched network, the response data causing the terminal device to present the post-processed natural language text in an editable region of a user interface.
