Electronic Arts Inc.

Electronics Arts, Inc.

Harold Henry Chaput

Navid Aghdaie

Kazi Atif-Uz Zaman

Mohsen Sardari

Kenneth Alan Moss

John Kolen

Wolfram Sebastian Starke

Yiwei Zhao

Igor Borovikov

Kazi Zaman

Han Liu

Kilol Gupta

Zahra Shakeri

Wei Wang

Jesse Hans Stokes Harder

Zebin Chen

Nitish Victor

Daniel Lee Kading

John Christopher Epler

Graham Francis Scott

Thomas Murray Perlinski

Sundeep Narravula

Meng Wu

Pawel Piotr Wrotek

Mattias Teye

Tushar Agarwal

Jervis Pinto

Jingwen Liang

Yunqi Zhao

Qilian Yu

Ulf Martin Lucas Singh-Blom

Ulf Erik Kristoffer Sjöö

Hitoshi Nishimura

Caedmon Somers

Jason Rupert

Dhaval Shah

David Auclair

Asim Shahjahan

Jesse Harder

Ahmad Beirami

Mathieu Guindon

Ian Szufnara

Fraser Lott

Wu Hai

Wilson Venhola

Zhaoming Xie

Maziar Sanjabi

Mihai Anghelescu

Yichi Zhang

Yangming Wen

Gordon Durity

A video game system and method analyze virtual contact between an avatar and a virtual object within a video game. The point of contact of the virtual contact on the virtual object and/or the intensity of contact of the virtual contact may then be used to determine a subsequent virtual action to be performed within the video game. The virtual action, with any virtual movement thereof, may be carried out in a realistic manner within the video game by determining a virtual trajectory of the motion. The virtual trajectory may be determined using a motion model. The motion model may provide the virtual trajectory of the virtual object based at least in part on one or more parameters of the virtual object, such as a weight parameter. The motion model may be trained using training video clips with realistic motion of virtual objects.

Object motion within a video game

Embodiments of the systems and methods described herein provide a terrain generation and population system that can determine terrain population rules for terrain population objects and features when placing objects and features in a three dimensional virtual space. As such, the terrain generation and population system can generate realistic terrain for use in game. The terrain generation and population system can receive an image, such as a satellite image, and utilize artificial intelligence to perform image segmentation at the pixel level to segment features and/or objects in the image. The game terrain system can automatically detect and apply feature and object masks based on the identified features and/or objects from the image segmentation. The game terrain system can place the features and/or objects in corresponding masks in the three dimensional space according to the application of terrain population rules.

Terrain generation and population system

In some embodiments, the dynamic animation generation system can provide a deep learning framework to produce a large variety of martial arts movements in a controllable manner from unstructured motion capture data. The system can imitate animation layering using neural networks with the aim to overcome challenges when mixing, blending and editing movements from unaligned motion sources. The system can synthesize movements from given reference motions and simple user controls, and generate unseen sequences of locomotion, but also reconstruct signature motions of different fighters. For achieving this task, the dynamic animation generation system can adopt a modular framework that is composed of the motion generator, that maps the trajectories of a number of key joints and root trajectory to the full body motion, and a set of different control modules that map the user inputs to such trajectories.

Neural animation layering for synthesizing movement

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.

Dynamic modifications of single player and multiplayer mode in a video game

The present disclosure provides a periodic autoencoder that can be used to generate a general motion manifold structure using local periodicity of the movement whose parameters are composed of phase, frequency, and amplitude. The periodic autoencoder is a novel neural network architecture that can learn periodic features from large unstructured motion datasets in an unsupervised manner. The character movements can be decomposed into multiple latent channels that can capture the non-linear periodicity of different body segments during synchronous, asynchronous, and transition movements while progressing forward in time, such that it captures spatial data and temporal data associated with the movements.

Learning character motion alignment with periodic autoencoders

Systems and methods for identifying one or more facial expression parameters associated with a pose of a character are disclosed. A system may execute a game development application to identify facial expression parameters for a particular pose of a character. The system may receive an input identifying the pose of the character. Further, the system may provide the input to a machine learning model. The machine learning model may be trained based on a plurality of poses and expected facial expression parameters for each pose. Further, the machine learning model can identify a latent representation of the input. Based on the latent representation of the input, the machine learning model can generate one or more facial expression parameters of the character and output the one or more facial expression parameters. The system may also generate a facial expression of the character and output the facial expression.

Predicting facial expressions using character motion states

Use of pose prediction models enables runtime animation to be generated for an electronic game. The pose prediction model can predict a character pose of a character based on joint data for a pose of the character in a previous frame. Further, by using environment data, it is possible to modify the prediction of the character pose based on a particular environment in which the character is location. Advantageously, the use of machine learning enables prediction of character movement in environment in which it is difficult or impossible to obtain motion capture data.

Dynamic locomotion adaptation in runtime generated environments

This specification describes a computer-implemented method of generating context-dependent speech audio in a video game. The method comprises obtaining contextual information relating to a state of the video game. The contextual information is inputted into a prosody prediction module. The prosody prediction module comprises a trained machine learning model which is configured to generate predicted prosodic features based on the contextual information. Input data comprising the predicted prosodic features and speech content data associated with the state of the video game is inputted into a speech audio generation module. An encoded representation of the speech content data dependent on the predicted prosodic features is generated using one or more encoders of the speech audio generation module. Context-dependent speech audio is generated, based on the encoded representation, using a decoder of the speech audio generation module.

Speech prosody prediction in video games

A method, computer-readable storage medium, and device for generating a master representation of input models. The method comprises: receiving a first base mesh and a second base mesh, wherein the first base mesh has a first topology and is associated with a first set of blendshapes to deform the first base mesh, the second base mesh has a second topology and is associated with a second set of blendshapes to deform the second base mesh, and the second topology is different from the first topology; combining the first topology and the second topology into a combined mesh topology representation; combining the first set of blendshapes and the second set of blendshapes into a combined blendshape representation; and outputting the combined mesh topology representation and the combined blendshape representation as a master representation, wherein the master representation can be queried with a target topology and blendshape.

Systems and methods for generating a model database with blendshape representation

Various aspects of the subject technology relate to systems, methods, and machine-readable media for generating insights for video games. The method includes gathering information regarding a player for a plurality of video games, the information comprising at least one of in-world state data, player action data, player progression data, and/or real-world events relevant to each video game. The method also includes tracking events in at least one video game of the plurality of video games, the events comprising an action event or a standby event. The method also includes determining that an event of the tracked events is an action event. The method also includes generating insights regarding the action event based on the information gathered regarding the player, the insights for improving the player's performance in the video game. The method also includes relaying the insights to the player to improve the player's performance in the video game.

Animated and personalized coach for video games

The present disclosure provides embodiments for joint twist generation for animation. The system can utilize a neural network, also referred to as a deep neural network, which utilizes machine learning processes in order to create animation data that are more life-like and realistic. The system can obtain a set of axis vectors for a rig of a virtual character model; obtain a twist model for the rig; input the set of axis vectors to the twist model to obtain a set of twist vectors; and determine animation data based on the set of axis vectors and the set of twist vectors.

Joint twist generation for animation

A system and method for providing live gameplay updates receives a modification for a videogame, the modification affecting a gameplay aspect of the videogame during execution of the videogame. The system and method determine a target group for deploying the modification. The target group includes first one or more live instances of the videogame. The system and method deploy the modification to the target group and receive gameplay data associated with the gameplay aspect from the target group. The system and method deploy the modification to second one or more live instances of the videogame based at least in part on an analysis of the received gameplay data.

Live gameplay updates

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.

Automated pipeline selection for synthesis of audio assets

Systems and methods are provided for enhanced animation generation based on generative control models. An example method includes accessing an autoencoder trained based on character control information generated using motion capture data, the character control information indicating, at least, trajectory information associated with the motion capture data, and the autoencoder being trained to reconstruct, via a latent feature space, the character control information. First character control information associated with a trajectory of an in-game character of an electronic game is obtained. A latent feature representation is generated and the latent feature representation is modified. A control signal is output to a motion prediction network for use in updating a character pose of the in-game character.

Enhanced animation generation based on generative control

A computing system may provide functionality for controlling an animated model to perform actions and to perform transitions therebetween. The system may determine, from among a plurality of edges from a first node of a control graph to respective other nodes of the control graph, a selected edge from the first control node to a selected node. The system may then determine controls for an animated model in a simulation based at least in part on the selected edge, control data associated with the selected node, a current simulation state of the simulation, and a machine learned algorithm, determine an updated simulation state of the simulation based at least in part on the controls for the animated model, and adapt one or more parameters of the machine learned algorithm based at least in part on the updated simulation state and a desired simulation state.

Learning character model animations with a layer-wise mixture-of-experts network

Growing community of inventors

Castro Valley, CA, United States of America

Harold Henry Chaput