# US 2026/0288856 A1 · System

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

- Number: US 2026/0288856 A1
- Document ID: US20260288856A1
- Status: Published application
- Kind: A1
- Application number: 19/564,474
- Filed: 2026-03-12
- Published: 2026-09-24
- Earliest claimed: 2025-03-19
- Assignees: SoftBank Group Corp.
- Applicants: SoftBank Group Corp.
- Inventors: Kohei HAGIHARA
- Web page: https://patentgrep.com/patent/US20260288856A1

## CPC

- G06F 16/435: Filtering based on additional data, e.g. user or group profiles
- G06F 16/438: Presentation of query results
- G06F 16/45: Clustering; Classification
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

A system includes a processor that is configured to: collect daily life data of a user through a sensor device and an application, and analyze the collected data to manage an amount of exercise of the user and meal content of the user; analyze voice data and text data of the user to recognize an emotion of the user and adjust content of advice and encouragement based on a recognition result, and generate a prompt for providing the advice and the encouragement to the user by using a generative artificial intelligence model; and generate, by using the prompt, a coach that imitates a person or a character preferred by the user.

## Claims

### Claim 1

A system comprising:
circuitry configured to:
receive, via a communication interface coupled to a packet-switched network, from a terminal device, time-series sensor data and multimodal input data associated with a user;
apply a first neural network model to the multimodal input data to generate a state classification label representing a detected condition of the user;
generate, based on the time-series sensor data and the state classification label, a numerical feature vector encoding a plurality of attributes of the user;
construct a natural-language prompt sentence incorporating at least a portion of the numerical feature vector and the state classification label;
execute a generative neural network model with the natural-language prompt sentence as an input to produce structured output data comprising an inference result and a persona-adapted output representation; and
transmit, via the communication interface coupled to the packet-switched network, the structured output data to the terminal device for presentation to the user.

### Claim 2 (depends on 1)

The system according to claim 1, wherein the circuitry is further configured to: receive, from the terminal device, feedback data indicating an evaluation of the structured output data by the user; and update at least one of the generative neural network model and a process of constructing the natural-language prompt sentence based on a correspondence among the feedback data, the numerical feature vector, and the natural-language prompt sentence.

### Claim 3 (depends on 2)

The system according to claim 2, wherein the circuitry is further configured to: aggregate a plurality of feedback data records associated with a plurality of numerical feature vectors; and modify weighting parameters used in constructing the natural-language prompt sentence based on correlation analysis between the plurality of feedback data records and corresponding state classification labels.

### Claim 4 (depends on 3)

The system according to claim 3, wherein: the multimodal input data comprises voice data captured by a microphone of the terminal device and text data input by the user; the state classification label represents an emotional state of the user; and the circuitry adjusts content of the structured output data based on the emotional state.

### Claim 5 (depends on 4)

The system according to claim 4, wherein the circuitry is further configured to:
apply a sentiment analysis function to the text data to extract a sentiment score; and
concatenate the sentiment score with the emotional state to produce an augmented state classification label used in constructing the natural-language prompt sentence.

### Claim 6 (depends on 1)

The system according to claim 1, wherein the circuitry is further configured to:
preprocess the time-series sensor data using a tabular data processing function to generate feature information; and apply a machine-learning feature transformation function to the feature information to produce the numerical feature vector, the machine-learning feature transformation function comprising at least one of encoding categorical variables, scaling numerical values, and transforming text-based fields into numerical representations.

### Claim 7 (depends on 6)

The system according to claim 6, wherein the circuitry is further configured to: apply dimensionality reduction to the numerical feature vector to produce a reduced-dimension feature vector; and use the reduced-dimension feature vector in constructing the natural-language prompt sentence.

### Claim 8 (depends on 7)

The system according to claim 7, wherein: the time-series sensor data comprises body activity data derived from an accelerometer of the terminal device and location data derived from a positioning receiver of the terminal device; and the feature information comprises aggregated values of body activity and amounts of location movement computed over a defined time window.

### Claim 9 (depends on 8)

The system according to claim 8, wherein: the body activity data represents an amount of exercise of the user including at least one of step count, distance traveled, and active minutes; and the circuitry computes lifestyle indices from the aggregated values of body activity and the amounts of location movement.

### Claim 10 (depends on 1)

The system according to claim 1, wherein the circuitry is further configured to: receive, from the terminal device, image data captured by a camera of the terminal device; and apply an image analysis neural network to the image data to generate classification labels and estimated quantity values associated with objects depicted in the image data.

### Claim 11 (depends on 10)

The system according to claim 10, wherein the circuitry is further configured to: receive transaction data from an external information processing apparatus, the transaction data comprising item identifiers, quantities, and timestamps; and correlate the classification labels with the transaction data using timestamp alignment and string similarity to produce refined estimation values.

### Claim 12 (depends on 11)

The system according to claim 11, wherein: the image data represents a meal of the user; the classification labels comprise food category labels; the refined estimation values comprise nutritional content values including calorie estimates and macronutrient ratios; and the circuitry incorporates the nutritional content values into the numerical feature vector.

### Claim 13 (depends on 1)

The system according to claim 1, wherein the persona-adapted output representation comprises: profile information defining characteristics of a virtual entity including at least one of a name, a personality description, a communication style, and a coaching style; and one or more guidance messages generated by the generative neural network model in accordance with the profile information.

### Claim 14 (depends on 13)

The system according to claim 13, wherein the circuitry is further configured to: construct a second natural-language prompt sentence that instructs the generative neural network model to reformat the inference result in a style defined by the profile information; and execute the generative neural network model with the second natural-language prompt sentence to produce the one or more guidance messages conforming to tone and length constraints specified in the second natural-language prompt sentence.

### Claim 15 (depends on 14)

The system according to claim 14, wherein: the profile information is generated based on a user preference setting indicating a preferred person or character; and the virtual entity imitates the preferred person or character in delivering the one or more guidance messages to the user as a coach providing advice and encouragement regarding exercise and diet.

### Claim 16 (depends on 1)

The system according to claim 1, wherein the circuitry is further configured to: parse the structured output data to extract labeled sections comprising at least a summary section, a risk identification section, and a recommendation section; and store the labeled sections in an advice log associated with the user and with the underlying time-series sensor data used to generate the structured output data.

### Claim 17 (depends on 16)

The system according to claim 16, wherein: the risk identification section identifies health-related concerns based on the numerical feature vector; the recommendation section comprises personalized lifestyle recommendations addressing the health-related concerns; and the circuitry transmits a notification payload to the terminal device indicating availability of updated advice content derived from meal content analysis, exercise tracking, and emotion recognition of the user.

### Claim 18

A system comprising:
circuitry configured to:
receive, via a communication interface coupled to a packet-switched network, from a terminal device, time-series sensor data and multimodal input data associated with a user;
apply a first neural network model to the multimodal input data to generate a state classification label;
preprocess the time-series sensor data using a tabular data processing function to generate feature information, and apply a machine-learning feature transformation function to the feature information to produce a numerical feature vector;
construct a natural-language prompt sentence incorporating the numerical feature vector and the state classification label, the natural-language prompt sentence comprising explicit instructions specifying a required output format and style constraints;
execute a generative neural network model with the natural-language prompt sentence as an input to produce structured output data comprising an inference result and a persona-adapted output representation including profile information and one or more guidance messages; and
transmit, via the communication interface coupled to the packet-switched network, the structured output data to the terminal device.

### Claim 19 (depends on 18)

The system according to claim 18, wherein the circuitry is further configured to: receive feedback data from the terminal device indicating an evaluation of at least one of the persona-adapted output representation and the one or more guidance messages; and perform feedback-aware fine-tuning of the generative neural network model by constructing training examples from the natural-language prompt sentence, the structured output data, and the feedback data, and applying a reward-weighted loss function during gradient-based optimization to adjust model parameters.

### Claim 20

A method performed by circuitry, the method comprising:
receiving, via a communication interface coupled to a packet-switched network, from a terminal device, time-series sensor data and multimodal input data associated with a user;
applying a first neural network model to the multimodal input data to generate a state classification label representing a detected condition of the user;
generating, based on the time-series sensor data and the state classification label, a numerical feature vector encoding a plurality of attributes of the user;
constructing a natural-language prompt sentence incorporating at least a portion of the numerical feature vector and the state classification label;
executing a generative neural network model with the natural-language prompt sentence as an input to produce structured output data comprising an inference result and a persona-adapted output representation; and
transmitting, via the communication interface coupled to the packet-switched network, the structured output data to the terminal device for presentation to the user.
