Abstract: Techniques are described for training neural networks on variable length datasets. The numeric representation of the length of each training sample is randomly perturbed to yield a pseudo-length, and the samples sorted by pseudo-length to achieve lower zero padding rate (ZPR) than completely randomized batching (thus saving computation time) yet higher randomness than strictly sorted batching (thus achieving better model performance than strictly sorted batching).

Abstract: A method for building an artificial intelligence (AI) model. The method includes accessing data related to monitored behavior of a user. The data is classified, wherein the classes include an objective data class identifying data relevant to a group of users including the user, and a subjective data class identifying data that is specific to the user. Objective data is accessed and relates to monitored behavior of a plurality of users including the user. The method includes providing as a first set of inputs into a deep learning engine performing AI the objective data and the subjective data of the user, and a plurality of objective data of the plurality of users. The method includes determining a plurality of learned patterns predicting user behavior when responding to the first set of inputs. The method includes building a local AI model of the user including the plurality of learned patterns.

Abstract: A method for building an artificial intelligence (AI) model. The method includes accessing data related to monitored behavior of a user. The data is classified, wherein the classes include an objective data class identifying data relevant to a group of users including the user, and a subjective data class identifying data that is specific to the user. Objective data is accessed and relates to monitored behavior of a plurality of users including the user. The method includes providing as a first set of inputs into a deep learning engine performing AI the objective data and the subjective data of the user, and a plurality of objective data of the plurality of users. The method includes determining a plurality of learned patterns predicting user behavior when responding to the first set of inputs. The method includes building a local AI model of the user including the plurality of learned patterns.

Abstract: Techniques are described for training neural networks on variable length datasets. The numeric representation of the length of each training sample is randomly perturbed to yield a pseudo-length, and the samples sorted by pseudo-length to achieve lower zero padding rate (ZPR) than completely randomized batching (thus saving computation time) yet higher randomness than strictly sorted batching (thus achieving better model performance than strictly sorted batching).

Abstract: A system enhances existing audio-visual content with an action a scene annotation module, an action description module, both of which are coupled to a controller. The scene annotation module classifies scene elements from an image frame received from a host system and generates a caption describing the scene elements. The scene annotation module includes a first neural network configured to generate a feature vector from the image frame and a second neural network configured to generate a caption describing elements within the image frame from the feature vector. The action description module recognizes action happening within one or more image frames received from the host system and generates a description of the action happening within one or more image frames.

Abstract: A system enhances existing audio-visual content with audio describing the setting of the visual content. A scene annotation module classifies scene elements from an image frame received from a host system and generates a caption describing the scene elements.

Abstract: Graphical style modification may be implemented using machine learning. A color accommodation module receives an image frame from a host system and generates a color-adapted version of the image frame. A Graphical Style Modification module generates a style adapted video stream by applying a style adapted from a target image frame to each image frame in a buffered video stream.

Abstract: Techniques are described for training neural networks on variable length datasets. The numeric representation of the length of each training sample is randomly perturbed to yield a pseudo-length, and the samples sorted by pseudo-length to achieve lower zero padding rate (ZPR) than completely randomized batching (thus saving computation time) yet higher randomness than strictly sorted batching (thus achieving better model performance than strictly sorted batching).

Abstract: Methods and systems are provided for displaying voice input of a spectator in a video game. The method includes receiving, by a server, the voice input produced by the spectator while viewing video game video of the video game. The method includes examining, by the server, the voice input to identify speech characteristics associated with the voice input of the spectator. The method includes processing, by the server, the voice input to generate a spectator video that includes text images representing the voice input of the spectator. In one embodiment, the text images are configured to be adjusted in visual appearance based on the speech characteristics of the voice input, wherein the text images are directed in a field of view of an avatar of the spectator. The method includes combining, by the server, video game video with the spectator video to produce an overlay of the text images graphically moving in toward a game scene provided by the video game video.

Abstract: Accommodation for color or visual impairments may be implemented by selective color substitution. A color accommodation module receives an image frame from a host system and generates a color-adapted version of the image frame. The color accommodation module may include a rule based filter that substitutes one or more colors within the image frame with one or more corresponding alternative colors.

Abstract: Techniques are described for training neural networks on variable length datasets. The numeric representation of the length of each training sample is randomly perturbed to yield a pseudo-length, and the samples sorted by pseudo-length to achieve lower zero padding rate (ZPR) than completely randomized batching (thus saving computation time) yet higher randomness than strictly sorted batching (thus achieving better model performance than strictly sorted batching).

Abstract: Methods and systems are provided for displaying voice input of a spectator in a video game. The method includes receiving, by a server, the voice input produced by the spectator while viewing video game video of the video game. The method includes examining, by the server, the voice input to identify speech characteristics associated with the voice input of the spectator. The method includes processing, by the server, the voice input to generate a spectator video that includes text images representing the voice input of the spectator. In one embodiment, the text images are configured to be adjusted in visual appearance based on the speech characteristics of the voice input, wherein the text images are directed in a field of view of an avatar of the spectator. The method includes combining, by the server, video game video with the spectator video to produce an overlay of the text images graphically moving in toward a game scene provided by the video game video.

Abstract: A chatbot learns a person's related “intents” when asking for information and thereafter, in response to an initial query, which the chatbot answers, the chatbot generates a secondary dialogue, either providing the person with additional information or inquiring as to whether the person wishes to know more about a subject. The chatbot may use an external trigger such as time, event, etc. and automatically generate a query or give information to the person without any initial query from the person.

Abstract: Graphical style modification may be implemented using machine learning. A color accommodation module receives an image frame from a host system and generates a color-adapted version of the image frame. A Graphical Style Modification module generates a style adapted video stream by applying a style adapted from a target image frame to each image frame in a buffered video stream.

Abstract: Graphical style modification may be implemented using machine learning. A color accommodation module receives an image frame from a host system and generates a color-adapted version of the image frame. A Graphical Style Modification module receives a first image frame from a host system and applies a style adapted from a second image frame to the first image frame to generate a style adapted first image frame.

Abstract: Accommodation for color or visual impairments may be implemented by selective color substitution. A color accommodation module receives an image frame from a host system and generates a color-adapted version of the image frame. The color accommodation module may include a rule based filter that substitutes one or more colors within the image frame with one or more corresponding alternative colors.

Abstract: A chatbot learns a person's related “intents” when asking for information and thereafter, in response to an initial query, which the chatbot answers, the chatbot generates a secondary dialogue, either providing the person with additional information or inquiring as to whether the person wishes to know more about a subject. The chatbot may use an external trigger such as time, event, etc. and automatically generate a query or give information to the person without any initial query from the person.

Abstract: Graphical style modification may be implemented using machine learning. A color accommodation module receives an image frame from a host system and generates a color-adapted version of the image frame. A Graphical Style Modification module receives a first image frame from a host system and applies a style adapted from a second image frame to the first image frame to generate a style adapted first image frame.

Abstract: A system enhances existing audio visual content with audio describing the action and setting of the visual content. The system may also provide subtitle content describing the important sound or sounds occurring within audio. Accommodation for color or visual impairments may be implemented by selective color substitution. A Graphical Style Modification module may apply a style from one image to another to adapt the style of a video per a gamer's preference.

Abstract: Accommodation for color or visual impairments may be implemented by selective color substitution. A color accommodation module receives an image frame from a host system and generates a color-adapted version of the image frame. The color accommodation module may include a rule based filter that substitutes one or more colors within the image frame with one or more corresponding alternative colors.