Back to results
US 2026/0284534 A1Published application
System
AssigneeSoftBank Group Corp.
InventorSeiya YAMAZAKI
Life of this patent
Estimating the term…
Abstract
A system includes a processor that is configured to use a prompt to instruct a generative AI model to transplant past match information and a persona of a real-life expert into the generative AI model, use a prompt to instruct initiation of a match based on user input so that a user can play a board game with an admired expert at any time, and use a prompt to instruct recognition of an emotional state of the user and adjustment of at least one of a difficulty level of the match and a persona of an opponent in accordance with the emotional state of the user.
Claims20
A system comprising:
- circuitry configured to:
- receive, via a communication interface coupled to a packet-switched network, competition information and behavioral characteristics associated with high-skill individuals, convert the competition information into structured data, and store the structured data in a storage device;
- train, based on the structured data, a generative neural network model that receives game state information and style information as input and outputs move distributions and evaluation values, and hold the trained generative neural network model in a state usable as an opponent;
- receive, from a terminal device via the communication interface, style instruction information including a prompt sentence and current game state information, convert the prompt sentence into numerical style information by text embedding processing, generate model input data by integrating the numerical style information with the current game state information, and execute inference processing using the generative neural network model with the model input data as input to generate a move among legal moves;
- transmit, to the terminal device via the communication interface, response data including the generated move and explanation text indicating a reason for selection of the generated move;
- and
- store user moves, model moves, prompt sentences, game state information, and game results as battle history data in the storage device, analyze the battle history data to classify each move based on changes in evaluation values, and update parameters of the generative neural network model based on the battle history data and classification results.
4 more elements
A system comprising:
- circuitry configured to:
- train, based on structured data derived from competition information and behavioral characteristics of high-skill individuals, a generative neural network model that receives game state information and style information as input and outputs move distributions and evaluation values;
- receive, from a terminal device via a communication interface coupled to a packet-switched network, a prompt sentence and current game state information, apply a text embedding model to the prompt sentence to generate a vector representation constituting numerical style information, and generate model input data by integrating the numerical style information with a feature vector derived from the current game state information;
- execute inference processing using the generative neural network model with the model input data as input to generate a move among legal moves, and transmit response data comprising the generated move and explanation text to the terminal device via the communication interface;
- and
- store user moves, model moves, prompt sentences, game state information, and game results as battle history data in a storage device, classify each stored move based on evaluation value changes using an evaluation function, and update model parameters of the generative neural network model based on the battle history data and classification results.
3 more elements
A method comprising:
- receiving, via a communication interface coupled to a packet-switched network, competition information and behavioral characteristics associated with high-skill individuals, converting the competition information into structured data, and storing the structured data in a storage device;
- training, based on the structured data, a generative neural network model that receives game state information and style information as input and outputs move distributions and evaluation values, and holding the trained generative neural network model in a state usable as an opponent;
- receiving, from a terminal device via the communication interface, style instruction information including a prompt sentence and current game state information, converting the prompt sentence into numerical style information by text embedding processing, generating model input data by integrating the numerical style information with the current game state information, and executing inference processing using the generative neural network model with the model input data as input to generate a move among legal moves;
- transmitting, to the terminal device via the communication interface, response data including the generated move and explanation text indicating a reason for selection of the generated move;
- and
- storing user moves, model moves, prompt sentences, game state information, and game results as battle history data in the storage device, analyzing the battle history data to classify each move based on changes in evaluation values, and updating parameters of the generative neural network model based on the battle history data and classification results.
3 more elements
Dependencies are inferred from the claim text.