Custom voice control of video game character

Patent metadata

Field	Value
Publication number	US-10926173-B2
Application number	US-201916436811-A
Country	US
Kind code	B2
Filing date	Jun 10, 2019
Priority date	Jun 10, 2019
Publication date	Feb 23, 2021
Grant date	Feb 23, 2021

Abstract

Systems and methods are disclosed for enabling a player of a video game to designate custom voice utterances to control an in-game character. One or more machine learning models may learn in-game character actions associated with each of a number of player-defined utterances based on player demonstration of desired character actions. During execution of an instance of a video game, current game state information may be provided to the one or more trained machine learning models based on an indication that a given utterance was spoken by the player. A system may then cause one or more in-game actions to be performed by a non-player character in the instance of the video game based on output of the one or more machine learning models.

First claim

What is claimed is:

1. A system comprising:
a data store that stores player interaction data associated with control by a player of one or more characters in a video game; and
a computing system in electronic communication with the data store and configured to execute computer-readable instructions that configure the computing system to:
retrieve player interaction data from the data store as at least a portion of training data for two or more machine learning models, wherein the training data identifies game information associated with each of a plurality of instances in which the player spoke a first utterance while controlling an in-game character in the video game, wherein the game information associated with each individual instance of the first utterance comprises at least (a) control input received, subsequent to the individual instance of the first utterance, from a control device of the player and (b) associated game state information when the first utterance was spoken in the individual instance;
train the two or more machine learning models to learn in-game character actions associated with the first utterance based on the training data, wherein the in-game character actions learned by the two or more machine learning models for the first utterance differ between at least two different game states, wherein the two or more machine learning models include a first model and a second model, wherein the first model is trained to associate an immediate in-game action with the first utterance, wherein the second model is trained to infer a goal associated with the first utterance;
subsequent to training the two or more machine learning models, execute an instance of the video game, wherein the player controls a first character in the instance of the video game using a first control device, wherein a second character in the instance of the video game is controlled by the system as a non-player character;
during execution of the instance of the video game, receive indication that the first utterance is spoken by the player;
based on the first utterance spoken by the player, provide at least current game state information as input to the two or more machine learning models;
determine at least one in-game action to be performed by the second character in the instance of the video game based on output of at least one of the two or more machine learning models; and
cause the second character in the instance of the video game to perform the at least one in-game action.

2. The system of claim 1 , wherein the at least one in-game action to be performed by the second character in the instance of the video game based on the output of the at least one of the two or more machine learning models comprises a series of actions to be performed by the second character to accomplish an inferred goal determined by the second model.

3. The system of claim 1 , wherein the second model is based at least in part on an inverse reinforcement learning algorithm.

4. The system of claim 1 , wherein the control device comprises at least one of a game controller, a keyboard or a mouse.

5. The system of claim 1 , wherein the instance of the video game includes a plurality of characters that are controlled by the system as non-player characters, and wherein the computing system is further configured to determine that the first utterance applies to the second character based on a portion of the utterance.

6. The system of claim 1 , wherein the second character is a team member of a team that includes the first character within the instance of the video game.

7. The system of claim 1 , wherein causing the second character in the instance of the video game to perform the at least one in-game action comprises causing the second character to behave similarly to behavior of a player-controlled character as captured in the training data.

8. A computer-implemented method comprising:
under the control of a computer system comprising computer hardware, the computer system configured with computer executable instructions:
obtaining training data for one or more machine learning models, wherein the training data comprises game information associated with each of a plurality of instances in which a player of a video game spoke at least a first utterance while controlling an in-game character in the video game, wherein the game information associated with each individual instance of the first utterance comprises at least one of control input received from a control device of the user or an in-game action performed by the in-game character subsequent to the individual instance of the first utterance, and wherein the game information further includes data representing a game state when the first utterance was spoken in the individual instance;
training the one or more machine learning models to learn in-game character actions associated with the first utterance based on the training data;
subsequent to training the one or more machine learning models, executing an instance of the video game, wherein the player controls a first character in the instance of the video game using a first control device, wherein a second character in the instance of the video game is a non-player character that is primarily controlled without player input;
during execution of the instance of the video game, receiving indication that the first utterance is spoken by the player;
based on the first utterance spoken by the player, providing at least current game state information as input to the one or more machine learning models;
determining at least one in-game action to be performed by the second character in the instance of the video game based on output of the one or more machine learning models;
causing the second character in the instance of the video game to perform the at least one in-game action;
receiving feedback from the player regarding whether the at least one in-game action performed by the second character was desired by the player when speaking the first utterance; and
updating at least one of the one or more machine learning models based on the feedback.

9. The computer-implemented method of claim 8 , further comprising executing a training environment within the video game, wherein the training environment presents simplified game states for player interaction, and wherein the training data is collected based on control input and utterances received from the player with the training environment.

10. The computer-implemented method of claim 8 , wherein the one or more machine learning models are trained to control one or more non-player characters in response to each of a plurality of utterances.

11. The computer-implemented method of claim 10 , wherein each of the plurality of utterances is one or more vocal sounds determined by the player and conveyed to the one or more machine learning models based in part on the player speaking the plurality of utterances during a training phase.

12. The computer-implemented method of claim 8 , wherein executing the instance of the video game comprising executing a hosted instance of the video game, wherein an application host system interacts with a game application executed on a computing device utilized by the player.

13. A non-transitory computer-readable storage medium having stored thereon computer-readable instructions that, when executed, configure a computing system to perform operations comprising:
training one or more machine learning models to learn in-game character actions associated with each of a plurality of player-defined utterances;
subsequent to training the one or more machine learning models, executing an instance of a video game, where

Assignees

• Electronic Arts Inc

Inventors

• Shah Dhaval Hemant
• Borovikov Igor
• Sardari Mohsen
• Kolen John
• Aghdaie Navid
• Zaman Kazi Atif-Uz

Classifications

• A63F13/424 Primary

  involving acoustic input signals, e.g. by using the results of pitch or rhythm extraction or voice recognition · CPC title

• A63F13/215 Primary

  comprising means for detecting acoustic signals, e.g. using a microphone · CPC title

• A63F13/63

  by the player, e.g. authoring using a level editor · CPC title