Abstract: This specification describes a computer-implemented method of generating speech audio for use in a video game, wherein the speech audio is generated using a voice convertor that has been trained to convert audio data for a source speaker into audio data for a target speaker. The method comprises receiving: (i) source speech audio, and (ii) a target speaker identifier. The source speech audio comprises speech content in the voice of a source speaker. Source acoustic features are determined for the source speech audio. A target speaker embedding associated with the target speaker identifier is generated as output of a speaker encoder of the voice convertor. The target speaker embedding and the source acoustic features are inputted into an acoustic feature encoder of the voice convertor. One or more acoustic feature encodings are generated as output of the acoustic feature encoder. The one or more acoustic feature encodings are derived from the target speaker embedding and the source acoustic features.

Abstract: An example method of automated selection of audio asset synthesizing pipelines includes: receiving an audio stream comprising human speech; determining one or more features of the audio stream; selecting, based on the one or more features of the audio stream, an audio asset synthesizing pipeline; training, using the audio stream, one or more audio asset synthesizing models implementing respective stages of the selected audio asset synthesizing pipeline; and responsive to determining that a quality metric of the audio asset synthesizing pipeline satisfies a predetermined quality condition, synthesizing one or more audio assets by the selected audio asset synthesizing pipeline.

Abstract: Systems and methods are provided for enhanced animation generation based on using motion mapping with local bone phases. An example method includes accessing first animation control information generated for a first frame of an electronic game including local bone phases representing phase information associated with contacts of a plurality of rigid bodies of an in-game character with an in-game environment. Executing a local motion matching process for each of the plurality of local bone phases and generating a second pose of the character model based on the plurality of matched local poses for a second frame of the electronic game.

Abstract: A computer-implemented method is provided of allowing a user to automatically transform domain knowledge into a machine learning model to be used in real-time operation of video games. The method comprises providing a user interface which allows a user to define domain knowledge relating to a video game by specifying one or more labeling functions; transforming the labeling functions into executable code; labeling raw data relating to the video game using the executable code to obtain labeled data; and applying an automated machine learning module to the labeled data to obtain a machine learning model.

Abstract: Systems and methods are provided for enhanced animation generation based on using motion mapping with local bone phases. An example method includes accessing first animation control information generated for a first frame of an electronic game including local bone phases representing phase information associated with contacts of a plurality of rigid bodies of an in-game character with an in-game environment. Executing a local motion matching process for each of the plurality of local bone phases and generating a second pose of the character model based on the plurality of matched local poses for a second frame of the electronic game.

Abstract: Embodiments of the systems and methods described herein provide a dynamic animation generation system that can apply a real-life video clip with a character in motion to a first neural network to receive rough motion data, such as pose information, for each of the frames of the video clip, and overlay the pose information on top of the video clip to generate a modified video clip. The system can identify a sliding window that includes a current frame, past frames, and future frames of the modified video clip, and apply the modified video clip to a second neural network to predict a next frame. The dynamic animation generation system can then move the sliding window to the next frame while including the predicted next frame, and apply the new sliding window to the second neural network to predict the following frame to the next frame.

Abstract: Embodiments of the systems and methods disclosed herein provide a request distribution system in which a request for resources may be executed by a plurality of workers. Upon receiving a request for resources from a user computing system, the request distribution system may select a subset of workers from the plurality of workers to execute the request within a time limit. Once the workers generate a plurality of outputs, each output associated with a quality level, the request distribution system may transmit the output associated with the highest quality level to the user computing system.

Abstract: This specification describes a system for generating positions of map items such as buildings, for placement on a virtual map. The system comprises: at least one processor; and a non-transitory computer-readable medium including executable instructions that when executed by the at least one processor cause the at least one processor to perform at least the following operations: receiving an input at a generator neural network trained for generating map item positions; generating, with the generator neural network, a probability of placing a map item for each subregion of a plurality of subregions of the region of the virtual map; and generating position data of map items for placement on the virtual map using the probability for each subregion.

Abstract: According to a first aspect of this specification, there is disclosed a computer implemented method comprising: training, based on an initial behavior goal and using reinforcement-learning, a reinforcement-learning model for controlling behavior of a non-playable character in a computer game environment; converting the trained reinforcement-learning model into a behavior tree model for controlling behavior of the non-playable character; editing, based on a user input, the behavior tree model to generate an updated behavior tree model for controlling behavior of the non-playable character; and outputting a final model for controlling non-player character behavior for use in the computer game environment, wherein the model for controlling non-player character behavior is based at least in part on the updated behavior tree model.

Abstract: An example method of automated selection of audio asset synthesizing pipelines includes: receiving an audio stream comprising human speech; determining one or more features of the audio stream; selecting, based on the one or more features of the audio stream, an audio asset synthesizing pipeline; training, using the audio stream, one or more audio asset synthesizing models implementing respective stages of the selected audio asset synthesizing pipeline; and responsive to determining that a quality metric of the audio asset synthesizing pipeline satisfies a predetermined quality condition, synthesizing one or more audio assets by the selected audio asset synthesizing pipeline.

Abstract: A collusion detection system may detect collusion between entities participating in online gaming. The collusion detection system may identify a plurality of entities associated with and opponents within an instance of an online game, determine social data associated with the plurality of entities, determine in-game behavior data associated with the plurality of entities, and determine, for one or more pairings of the plurality of entities, respective pairwise feature sets based at least in part on the social data and the in-game behavior data. The collusion detection system may then perform anomaly detection on the respective pairwise feature sets and, in response to the anomaly detection detecting one or more anomalous pairwise feature sets, output one or more suspect pairings of the plurality of entities corresponding to the one or more anomalous pairwise feature sets as suspected colluding pairings.

Abstract: Various aspects of the subject technology relate to systems, methods, and machine-readable media for user matchmaking. The method includes training a quality model and an embedding model based on historical data and user control options. The method also includes receiving user control options and matchmaking requests from users. The method also includes embedding, through the embedding model, user data regarding the users into an embedded space based on the received user control options and the matchmaking requests. The method also includes determining, based on the embedded user data, that a distance between two users satisfies a distance threshold. The method also includes matching the two users when the determined distance satisfies the distance threshold.

Abstract: Systems described herein may automatically and dynamically adjust the amount and type of computing resources usable to execute, process, or perform various tasks associated with a video game. Using one or more machine learning algorithms, a prediction model can be generated that uses the historical and/or current user interaction data obtained by monitoring the users playing the video game. Based on the historical and/or current user interaction data, future user interactions likely to be performed in the future can be predicted. Using the predictions of the users' future interactions, the amount and type of computing resources maintained in the systems can be adjusted such that a proper balance between reducing the consumption of computing resources and reducing the latency experienced by the users of the video game is achieved and maintained.

Abstract: This specification describes a system for generating positions of map items such as buildings, for placement on a virtual map. The system comprises: at least one processor; and a non-transitory computer-readable medium including executable instructions that when executed by the at least one processor cause the at least one processor to perform at least the following operations: receiving an input at a generator neural network trained for generating map item positions; generating, with the generator neural network, a probability of placing a map item for each subregion of a plurality of subregions of the region of the virtual map; and generating position data of map items for placement on the virtual map using the probability for each subregion.

Abstract: Various aspects of the subject technology relate to systems, methods, and machine-readable media for granting access to a game. The method includes receiving a request to access the game. The method also includes causing to display a set of challenge options representing challenges of varying difficulty levels for a user to select. The method also includes receiving a selection of a challenge option from a user device. The method also includes causing to present a challenge to the user. The method also includes receiving challenge data from the user including interactions with and results of the challenge. The method also includes determining whether the user passed the challenge based on the challenge data. The method also includes executing or enabling execution of the game upon determining that the user passed the challenge. The method also includes providing the user rewards associated with the game for passing the challenge.

Abstract: Embodiments of the systems and methods described herein can automatically measure the difficulty metrics associated with various aspects of a video game using an artificial intelligence system. The artificial intelligence system may include multiple game agents. Telemetry data associated with the gameplay of each game agent may be recorded while the game application is automatically executed by the game agents. The telemetry data may be communicated to a data analysis system which can calculate game difficulty metrics for various aspects of the game. The data analysis system can determine game difficulty associated with the various aspects based on the game difficulty metrics. The results from the data analysis system may be visualized and communicated to a game developer for updating the operations of the video game.

Abstract: Embodiments of the systems and methods disclosed herein provide a request distribution system in which a request for resources may be executed by a plurality of workers. Upon receiving a request for resources from a user computing system, the request distribution system may select a subset of workers from the plurality of workers to execute the request within a time limit. Once the workers generate a plurality of outputs, each output associated with a quality level, the request distribution system may transmit the output associated with the highest quality level to the user computing system.

Abstract: A method, computer-readable storage medium, and device for generating a character model. The method comprises: receiving an input image of a reference subject; processing the input image to generate a normalized image; identifying a set of features present in the normalized image, wherein each feature in the set of features corresponds to a portion of a head or body of the reference subject; for each feature in the set of features, processing at least a portion of the normalized image including the feature by a neural network model corresponding to the feature to generate a parameter vector corresponding to the feature; and combining the parameter vectors output by respective neural network models corresponding to respective features in the set of features to generate a parameterized character model corresponding to reference subject in the input image.

Abstract: A video game includes a single player mode where completion of storyline objectives advances the single player storyline. The video game also includes a multiplayer mode where a plurality of players can play on an instance of a multiplayer map. Storyline objectives from the single player mode are selected and made available for completion to players in the multiplayer mode, and the single player storylines can be advanced by players completing respective storyline objectives while playing in the multiplayer mode. Combinations of storyline objectives are selected from pending storyline objectives for players connecting to a multiplayer game for compatibility with multiplayer maps. Constraints can be used to determine compatibility.

Abstract: According to an aspect of this specification, there is described a computer implemented method comprising: receiving input data, the input data comprising data relating to a user of a computer game; generating, based on the input data, one or more candidate challenges for the computer game; determining, using a machine-learned model, whether each of the one or more of the candidate challenges satisfies a threshold condition, wherein the threshold condition is based on a target challenge difficultly; in response to a positive determination, outputting the one or more candidate challenges that satisfy the threshold condition for use in the computer game by the user.