Abstract: An apparatus for generating datasets for training machine learning models includes: a receiving unit configured to receive video data comprising sequential image frames; a storage unit configured to store a plurality of the sequential image frames; and a selecting unit configured to select, for a target image frame, a subset of stored image frames, the subset providing contextual data relating to the target image frame for the machine learning model. The selecting unit is configured to successively generate sampling values, wherein a difference between successive sampling values increases with each successively generated sampling value; and the selecting unit is configured to select a given image frame from the stored image frames in dependence upon whether a number of sequential image frames between the given image frame and the target image frame coincides with one of the successively generated sampling values.

Abstract: System, process and device configurations are provided for analyzing and modifying a game environment. A method can include generating at least one game scenario for an electronic game including control settings for an audio configuration, a video configuration and a gameplay configuration. The method may include outputting the at least one game scenario using the control settings, and detecting user interest for the at least one game scenario. User interest for the at least one game scenario may be detected using a kernel function including a kernel element for each of a visual style, an audio style and a gameplay style. User profile data may be updated for the at least one user based on detected user interest for the at least one game scenario. Output of a game state may be controlled for the electronic game using the user profile.

Abstract: Techniques are provided to detect the motion of a player of a computer game and both convert images of the motion to text to allow another player to attempt to mimic it, with a multimodal model judging how well the second player mimics the first, and to map the player motion onto a player character (PC) in the game.

Abstract: A swarm of drones over a large open area or stadium project portions of computer game video onto the ground or a wall in the area. The drones are controlled by a hub to present a game played on phone or controller to gamify any space.

Abstract: A computer game being played in a parked vehicle can send signals to one or more vehicle actuators in synchronization with the game so that haptics can be generated by the vehicle's actuators as part of game play.

Abstract: Computer vision of a large audience in an arena or stadium with a massive display (such as an arena within a hemispherical display) is fed into a machine learning (ML) model to generate game play signals based on, e.g., the number of raised hands, synchronicity of a “wave”, areas of screen focus, jumping, leaning, and high-fiving.

Abstract: A data processing apparatus comprising circuitry configured to: execute a machine learning, ML, model configured to receive, as an input, a game state of a video game and first virtual physiological data indicative of a first virtual physiological state of an agent of the video game, and generate, as an output, a probability of each of a plurality of actions of the agent and second virtual physiological data indicative of a second, subsequent, virtual physiological state of the agent; and perform reinforcement learning to generate a policy for completion of a task by the agent, the reinforcement learning comprising, for each of a plurality of attempts at the task by the agent, executing one or more successive iterations of the ML model and, for each attempt, controlling the agent to perform a different respective set of actions based on the output probability of each of the plurality of actions at each of the one or more successive iterations.

Abstract: Exemplary embodiments include reinforcement learning systems and methods for discovering new techniques for playing a game. An exemplary system comprises: A determining agent configured to search at least one Internet platform for data related to a game scenario and determine at least one reward framework based on results from the search of the Internet platform, the reward framework being determined by at least one metric for a characteristic of the game scenario; and a reinforcement learning agent configured to perform a training and exploration loop comprising a plurality of iterations, each iteration comprising: playing at least one game scenario within a game by taking sequential in-game actions available in the game scenario; transmitting results of each of the plurality of sequential in-game actions to the determining agent; and receiving a reward for successful progression through the game scenario according to the reward framework determined by the determining agent.

Abstract: A data processing apparatus comprising circuitry configured to: obtain gameplay data indicating in-game behaviour of a human video game player; configure, using the obtained gameplay data, an artificial intelligence, AI, agent representing the human video game player; execute one or more AI agent video game sessions including the AI agent and obtain AI agent performance data indicating in-game performance of the AI agent during the one or more AI agent video game sessions; and perform matchmaking of the human video game player with another human video game player based on the obtained AI agent performance data.

Abstract: A method of generating a training system comprising, for a set of states and corresponding actions, training separate auto encoders; using interim encoded representations from each trained auto encoder as input to a machine learning recoder, wherein each recoder is trained with a respective multi-part loss function that discriminates output representations for different respective inputs within each recoder whilst converging representations for parallel current state and action inputs between the recoders.

Abstract: A data processing apparatus for detecting one or more errors for one or more video images for a video game, the data processing apparatus comprising receiving circuitry to receive a sequence of video images for the video game and one or more action inputs associated with the sequence of video images, prediction circuitry to generate a predicted video image in dependence on at least one video image of the sequence of video images and an action input associated with the at least one video image, the prediction circuitry comprising a trained machine learning model to generate the predicted video image, and error detection circuitry to detect, for one or more video images subsequent to the at least one video image, one or more errors in dependence on the predicted video image.

Abstract: A method of training an autonomous agent is provided, the method comprising: providing videogame data generated by a human playing a videogame as input data to a training network for training an autonomous agent for playing the videogame; generating videogame data of the trained autonomous agent playing the videogame; providing the videogame data of the trained autonomous agent playing the videogame data to a discriminator of a generative adversarial network ‘GAN’, the discriminator of the GAN being trained to distinguish videogame data of a human playing the videogame and videogame data of an autonomous agent playing the videogame; generating a classification, by the discriminator of the GAN, of the output videogame data of the trained autonomous agent playing the videogame as human or agent; and updating at least one of the training network and the discriminator of the GAN based on the classification generated by the discriminator of the GAN.

Abstract: A data processing apparatus comprises a captioning model to receive gameplay telemetry data indicative of one or more in-game properties for a session of a video game, the captioning model comprising an artificial neural network (ANN) trained to output caption data comprising one or more captions in dependence upon a learned mapping between gameplay telemetry data and caption data, one or more of the captions comprising one or more words for providing a visual description for the session of the video game, and output circuitry to output one or more of the captions.

Abstract: A decoder apparatus comprises receiving circuitry to receive caption data indicative of a language-based description for a first image and encoded data representative of the first image; and decoder circuitry comprising one or more trained machine learning models operable to generate a reconstructed image in dependence on the caption data and the encoded data, the reconstructed image having a higher image quality than an image quality associated with the encoded data representative of the first image.

Abstract: Systems and methods are disclosed for expanding the output modalities of a gameplay session. Gameplay input data, such as audio data, video data, haptics data, and Gameplay context data, can be processed using machine learning (ML) models to generate output state vectors, which can be used to drive state changes in remote devices, such as smarthome devices (e.g., smart lightbulbs, smart light switches, smart fans, smart thermostats, smart blinds, smart speakers, etc.). To generate the output state vectors, first gameplay features are extracted from the gameplay input data, then the gameplay features are processed by a state vector generation model to generate the state vectors. Each modality of gameplay input data can be processed by a corresponding feature extractor. Gameplay sessions can be evaluated to train the ML models.

Abstract: System, process and device configurations are provided for analyzing and modifying a game environment. A method can include generating at least one game scenario for an electronic game including control settings for an audio configuration, a video configuration and a gameplay configuration. The method may include outputting the at least one game scenario using the control settings, and detecting user interest for the at least one game scenario. User interest for the at least one game scenario may be detected using a kernel function including a kernel element for each of a visual style, an audio style and a gameplay style. User profile data may be updated for the at least one user based on detected user interest for the at least one game scenario. Output of a game state may be controlled for the electronic game using the user profile.

Abstract: A data processing apparatus for determining description data for describing content includes: a video captioning model to receive an input comprising at least video images associated with the content, wherein the video captioning model is trained to detect one or more predetermined motions of one or more animated objects in the video images and determine one or more captions in dependence on one or more of the predetermined motions, one or more of the captions comprising respective caption data comprising one or more words for describing one or more of the predetermined motions, the respective caption data comprising one or more of audio data, text data and image data; and output circuitry to output description data in dependence on one or more of the captions.

Abstract: An apparatus for generating datasets for training machine learning models includes: a receiving unit configured to receive video data comprising sequential image frames; a storage unit configured to store a plurality of the sequential image frames; and a selecting unit configured to select, for a target image frame, a subset of stored image frames, the subset providing contextual data relating to the target image frame for the machine learning model. The selecting unit is configured to successively generate sampling values, wherein a difference between successive sampling values increases with each successively generated sampling value; and the selecting unit is configured to select a given image frame from the stored image frames in dependence upon whether a number of sequential image frames between the given image frame and the target image frame coincides with one of the successively generated sampling values.