Abstract: A visual anomaly detection system can test a video game under test. The testing can involve applying captured video frames to a large language model using dynamically generated prompts. The captured video frames can be obtained directly from an output port of a user computing system enabling the video frames to be applied to the machine learning model without modification and with minimal to no user intervention. Additionally, the systems disclosed herein can control the user computing system hosting the video game under test enabling the test system to react to test results in real-time or near real-time (e.g., within milliseconds, while the video game is executing, before a next action is performed with respect to the video game, and the like) and to modify the testing process as tests are being performed.

Abstract: A gaming system may provide for interactable environment geometry (IEG) detection. The gaming system may determine a sensory perspective of a sense of a non-player character (NPC) in a virtual environment of a game simulation, generate perception data of the NPC from the sensory perspective of the sense, input the perception data of the NPC into a detection model associated with the sense and receive, from the detection model, detection data for a detected item. The gaming system may then generate an awareness-based character decision for the NPC based on the detection data of the detected item.

Abstract: A method of generating a three-dimensional (3D) model includes obtaining a set of two-dimensional (2D) images of a scene acquired by one or more cameras from a set of camera angles and camera positions, obtaining the set of camera angles and camera positions based on obtaining, for each 2D image in the set of 2D images, a respective camera angle and a respective camera position for the 2D image, training a neural radiance field (NeRF) model, using the set of 2D images and the set of camera angles and camera positions as a training dataset, to obtain a trained NeRF model, generating a set of 2D depth maps based on the trained NeRF model, and generating a 3D polygonal mesh representing at least one object of one or more objects in the scene based on the set of 2D depth maps.

Abstract: A spectator system may provide for spectating in online gaming. The spectator system may receive, at a spectator server, game state data from a game simulation server hosting an online game for one or more players, generate one or more spectator game state data corresponding to one or more spectator devices and output the one or more spectator game state data to the spectator devices. The spectator server may further output the game state data to another spectator server.

Abstract: Systems and methods are provided for enhanced face shape generation for archetypes of virtual entities based on generative modeling techniques. An example method includes training models based on synthetically generated faces and information associated with an authoring system. The modeling system being trained to reconstruct face shapes for virtual entities based on a latent space embedding of a face identity associated with specific archetypes.

Abstract: A quality analysis tool for visual-programming scripting languages uses machine learning to process changes from visual-programming environments. The quality analysis tool can receive data associated with a code submission via a visual-programming scripting language, process the data to identify features in the data that correspond to previously identified defects, apply a pattern matching algorithm to the identified features, determine a risk prediction based on a learned pattern recognition model associated with a pattern in the features, and transmit a notice of predicted risk. The quality analysis tool can train models for use with visual-programming scripting languages and visual-programming environments.

Abstract: This specification describes a computer-implemented method of generating expressive speech audio for use in a video game. The method comprises receiving user-input text data and receiving prosody data obtained from user-input speech style data. The user-input text data is processed by a text encoder to generate a text feature embedding. A speech feature embedding is generated by an embedding converter based upon the text feature embedding. The prosody data is processing by a prosody encoder to generate a prosody feature embedding. A speech signal is generated by a speech decoder based upon the speech feature embedding and the prosody feature embedding.

Abstract: Implementations described herein relate to a method for predicting motion capture data implemented by a video game development system comprising one or more processors, the method including: obtaining, by one or more of the processors of the video game development system, motion capture data; the motion capture data including one or more missing values; generating, by one or more of the processors of the video game development system, a predicted value for the one or more missing values, generating a predicted value includes processing the motion capture data using one or more machine learning models to determine the predicted value.

Abstract: A party system of a gaming environment provides to users of the gaming environment parties of sizes that extend beyond the party size limitations of a video game. An extendable party can dynamically create subset parties of users for gameplay as needed, without requiring the party to be ceased. Therefore, an extendable party allows a subset of users, or multiples thereof, to enter into gameplay while maintaining the party as a whole. A party system, video game, or gaming environment can be configured to apply one or more rules or policies to a party that limit or alter a feature or function of an extended party, such as to prevent players among extended parties from providing or receiving competitive advantages.

Abstract: Implementations disclosed herein are directed to a method for predicting motion capture data implemented by a video game development system including one or more processors. The method can include: obtaining, by one or more of the processors of the video game development system, motion capture data; the motion capture data can include one or more missing values; generating, by one or more of the processors of the video game development system, a predicted value for the one or more missing values. Generating a predicted value can include processing the motion capture data using one or more machine learning models to determine the predicted value.

Abstract: A system may perform motion capture using motion capture targets with concave reflector structures. For example, the motion capture target include a target body and a plurality of tracking markers located on respective portions of the surface of the target body. At least one tracking marker of the plurality of tracking markers may be a concave reflector structure including a tapered hole in the surface of the target body and at least a portion of a surface of the tapered hole may be reflective.

Abstract: A method of generating a three-dimensional (3D) model includes obtaining a set of two-dimensional (2D) images of a scene acquired by one or more cameras from a plurality of camera angles at a plurality of camera positions. Each 2D image corresponds to a respective camera angle and a respective camera position. The method further includes obtaining the respective camera angle and the respective camera position for each 2D image, and generating one or more semantic masks from the set of 2D images. Each semantic mask corresponds to a class of one or more objects in the scene. The method further includes training a neural radiance field (NeRF) model, using the set of 2D images and the one or more semantic masks as a training dataset, to obtain a trained NeRF model. The trained NeRF model is an implicit 3D model of the one or more objects in the scene.

Abstract: The present disclosure provides a system for customizing virtual entities via a multi-track system. The system can generate a user interface that uses multiple tracks to manage and display virtual entities and virtual objects. The virtual entity can be in-line with a track, and virtual display objects can move along tracks that intersect with the virtual entity track. When a particular virtual display object intersects with the virtual entity track, a three-dimensional representation of an item associated with the particular virtual display object can be rendered with the virtual entity.

Abstract: A method, device, and computer-readable storage medium for simplifying a mesh including bevels. The method includes: receiving a polygonal mesh representing a three-dimensional (3D) object; identifying a set of edges in the polygonal mesh as bevel edges; performing a mesh simplification operation on the polygonal mesh to generate a simplified mesh, wherein the mesh simplification operation removes at least one edge that includes a vertex that is of a bevel edge, and wherein two vertices in the polygonal mesh are collapsed to a collapse vertex in the simplified mesh; and updating stored normals of the collapse vertex based on copying stored normals of the two vertices removed from the polygonal mesh to the collapse vertex.

Abstract: Systems and methods for user interface navigation may include a computing system which causes display of a user interface including a set of first user interface elements. The computing system receives a request to expand a switcher menu including a set of second user interface elements including at least one of the first user interface elements. Each second user interface element, when selected within the switcher menu, is displayed with one or more respective third user interface elements. The computing system causes display of a set of third user interface elements corresponding to one of the second user interface elements, and causes, responsive to a second request to navigate from the switcher menu to one of the set of third user interface elements, the switcher menu to collapse and display the second user interface element with at least some of the third user interface elements.

Abstract: System and methods for using a deep learning framework to customize animation of an in-game character of a video game. The system can be preconfigured with animation rule sets corresponding to various animations. Each animation can be comprised of a series of distinct poses that collectively form the particular animation. The system can provide an animation-editing interface that enables a user of the video game to make modifications to at least one pose or frame of the animation. The system can realistically extrapolate these modifications across some or all portions of the animation. In addition or alternatively, the system can realistically extrapolate the modifications across other types of animations.

Abstract: A first set of contact points for a hull object colliding with a first primitive of a mesh object is determined. An edge clip is generated for an edge of the first primitive based on determining that the hull object generates contact points with the edge. A face walk is performed into a second primitive of the mesh object across the edge for the edge clip. A set of initial contact points for the hull object colliding with the second primitive is generated using the algorithm. A second set of contact points is generated based on projecting each initial contact point in a direction of a contact normal for the first set of contact points onto a plane created by extending a feature of the first primitive. Contact points in the first and second set of contact points are output, but not the contacts along the edge clip.

Abstract: A method, device, and computer-readable storage medium for generating levels-of-detail (LODs) for blendshape assets is disclosed. Given a base target and a set of blendshape targets, a LOD for the base target is generated based on performing edge collapse operations on the base target, where the edge collapse operations that are performed are informed by not only the geometry of the base target, but also the geometries of the blendshape targets. In some cases, the same edge collapse operations can also be performed on the blendshape targets to generate LODs for the blendshape targets, thereby maintaining topological consistency between the LODs for the base target and the blendshape targets. In some cases, a morphing operator can take as input a LOD of the base target and generate a LOD for a blendshape target.