Back to results
US 2026/0289183 A1Published application
System
AssigneeSoftBank Group Corp.
InventorMikiko TAKEUCHI
Life of this patent
Estimating the term…
Abstract
A system includes a processor that is configured to receive information via an interface that receives information from a user, analyze the received information by using a generative AI model and evaluate legal concerns, security impacts, and customer reactions, generate a prompt for sorting the information to appropriate departments based on a result of the analysis, use an emotion analysis technique to analyze an emotional state of the user, and generate a prompt for setting a priority of the information based on the emotional state.
Claims20
A system comprising:
- circuitry configured to:
- receive natural language input data from a terminal;
- perform natural language preprocessing on the natural language input data by applying character string normalization, tokenization, sentence segmentation, and language identification to generate analysis data;
- apply a multi-label classification model to the analysis data to extract terms and classify the extracted terms into at least a regulation category, a technology category, and a market category to generate structured category data;
- generate a first prompt sentence by serializing an analysis result derived from the structured category data into a structured prompt template that encodes the regulation category terms, the technology category terms, the market category terms, and section-specific instruction tokens constraining output scope;
- input the first prompt sentence to a generative artificial intelligence model to generate a structured document comprising at least a legal concerns section, a security impacts section, a customer reactions section, and a risk reduction measures section;
- generate a second prompt sentence encoding a multi-dimensional feature vector derived from the legal concerns section, the security impacts section, and the customer reactions section;
- and
- input the second prompt sentence to the generative artificial intelligence model to generate an organizational assignment result identifying processing units corresponding to the multi-dimensional feature vector.
6 more elements
A system comprising:
- circuitry configured to:
- receive natural language input data from a terminal and perform natural language preprocessing including tokenization, sentence segmentation, and language identification to generate analysis data;
- apply a multi-label classification model to the analysis data to classify extracted terms into at least a regulation category, a technology category, and a market category to generate structured category data;
- generate a first prompt sentence serializing an analysis result derived from the structured category data and input the first prompt sentence to a generative artificial intelligence model to generate a structured document having predefined sections;
- apply an emotion analysis technique to input information from the terminal to calculate an emotional state and generate a priority prompt sentence based on the emotional state and the structured category data;
- and
- input the priority prompt sentence to the generative artificial intelligence model to control at least one of an allocation order and a degree of emphasis in document content according to a determined priority.
4 more elements
A method performed by circuitry, the method comprising:
- receiving natural language input data from a terminal;
- performing natural language preprocessing on the natural language input data by applying character string normalization, tokenization, sentence segmentation, and language identification to generate analysis data;
- applying a multi-label classification model to the analysis data to extract terms and classify the extracted terms into at least a regulation category, a technology category, and a market category to generate structured category data;
- generating a first prompt sentence by serializing an analysis result derived from the structured category data into a structured prompt template encoding section-specific instruction tokens;
- inputting the first prompt sentence to a generative artificial intelligence model to generate a structured document comprising at least a legal concerns section, a security impacts section, and a customer reactions section;
- applying an emotion analysis technique to input information to calculate an emotional state and generating a priority prompt sentence based on the emotional state and the structured category data;
- and
- inputting the priority prompt sentence to the generative artificial intelligence model to control at least one of an allocation order and a degree of emphasis in document content.
5 more elements
Dependencies are inferred from the claim text.