Abstract: A method for integration of real-world content into a game is described. The method includes receiving a request to play the game and accessing overlay multimodal data generated from a portion of real-world multimodal data received as user generated content (RGC). The overlay multimodal data relates to authored multimodal data generated for the game. The method includes replacing the authored multimodal data in one or more scenes of the game with the overlay multimodal data.

Abstract: Methods and systems for engaging an AI player of a user to play a video game on behalf of the user includes creating the AI player for the user using at least some of the attributes of the user, training the AI player using inputs provided by the user during game play of the video game, and providing access to the video game for game play to the AI player. The access allows the AI player to provide inputs to the video game that substantially mimics a play style of the user. Control of the game play of the video game can be transitioned to the user at any time during the game play of the AI player. The user can also control the game play of the AI player from a video recording of the game play.

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 wake-up word for a digital assistant may be specified by a user to trigger the digital assistant to respond to the wake-up word, with the user providing one or more initial pronunciations of the wake-up word. The wake-up word may be unique, or at least not determined beforehand by a device manufacturer or developer of the digital assistant. The initial pronunciation(s) of the keyword may then be augmented with other potential pronunciations of the wake-up word that might be provided in the future, and those other potential pronunciations may then be pruned down to a threshold number of other potential pronunciations. One or more recordings of the initial pronunciation(s) of the wake-up may then be used to train a phoneme recognizer model to better recognize future instances of the wake-up word being spoken by the user or another person using the initial pronunciation or other potential pronunciations.

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: Text and speech from a computer simulation are processed by a machine learning engine to animate the face of a computer avatar.

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: An inexperienced computer simulation player is assisted in playing the simulation by identifying in previously-played simulations successful players in terms of simulation play. The virtual weapons selected by those players are identified and a recommendation of the weapon presented to the inexperienced player of the computer simulation.

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 wake-up word for a digital assistant may be specified by a user to trigger the digital assistant to respond to the wake-up word, with the user providing one or more initial pronunciations of the wake-up word. The wake-up word may be unique, or at least not determined beforehand by a device manufacturer or developer of the digital assistant. The initial pronunciation(s) of the keyword may then be augmented with other potential pronunciations of the wake-up word that might be provided in the future, and those other potential pronunciations may then be pruned down to a threshold number of other potential pronunciations. One or more recordings of the initial pronunciation(s) of the wake-up may then be used to train a phoneme recognizer model to better recognize future instances of the wake-up word being spoken by the user or another person using the initial pronunciation or other potential pronunciations.

Abstract: Video and audio from a computer simulation are processed by a machine learning engine to identify candidate segments of the simulation for use in a video summary of the simulation. Text input is then used to reinforce whether a candidate segment should be included in the video summary.

Abstract: Text and speech from a computer simulation are processed by a machine learning engine to animate the face of a computer avatar.

Abstract: Video and audio from a computer simulation are processed by a machine learning engine to identify candidate segments of the simulation for use in a video summary of the simulation. Text input is then used to reinforce whether a candidate segment should be included in the video summary. Metadata can be added to the summary showing game summary information.

Abstract: A wake-up word for a digital assistant may be specified by a user to trigger the digital assistant to respond to the wake-up word, with the user providing one or more initial pronunciations of the wake-up word. The wake-up word may be unique, or at least not determined beforehand by a device manufacturer or developer of the digital assistant. The initial pronunciation(s) of the keyword may then be augmented with other potential pronunciations of the wake-up word that might be provided in the future, and those other potential pronunciations may then be pruned down to a threshold number of other potential pronunciations. One or more recordings of the initial pronunciation(s) of the wake-up may then be used to train a phoneme recognizer model to better recognize future instances of the wake-up word being spoken by the user or another person using the initial pronunciation or other potential pronunciations.

Abstract: An inexperienced computer simulation player is assisted in playing the simulation by identifying in previously-played simulations successful players in terms of simulation play. The virtual weapons selected by those players are identified and a recommendation of the weapon presented to the inexperienced player of the computer simulation.

Abstract: A wake-up word for a digital assistant may be specified by a user to trigger the digital assistant to respond to the wake-up word, with the user providing one or more initial pronunciations of the wake-up word. The wake-up word may be unique, or at least not determined beforehand by a device manufacturer or developer of the digital assistant. The initial pronunciation(s) of the keyword may then be augmented with other potential pronunciations of the wake-up word that might be provided in the future, and those other potential pronunciations may then be pruned down to a threshold number of other potential pronunciations. One or more recordings of the initial pronunciation(s) of the wake-up may then be used to train a phoneme recognizer model to better recognize future instances of the wake-up word being spoken by the user or another person using the initial pronunciation or other potential pronunciations.

Abstract: An arrangement and corresponding method are described for multi-domain natural language processing. Multiple parallel domain pipelines are used for processing a natural language input. Each domain pipeline represents a different specific subject domain of related concepts. Each domain pipeline includes a mention module that processes the natural language input using natural language understanding (NLU) to determine a corresponding list of mentions, and an interpretation generator that receives the list of mentions and produces a rank-ordered domain output set of sentence-level interpretation candidates. A global evidence ranker receives the domain output sets from the domain pipelines and produces an overall rank-ordered final output set of sentence-level interpretations.

Abstract: A computer-implemented method for dynamically presenting a prewritten text in a graphical user interface is disclosed. The method comprises receiving a text artifact, storing the text artifact in a memory device of a computer, retrieving the text artifact, displaying the text artifact on the display screen of the computer, receiving a vocal input, generating a text file representing the words spoken in the vocal input, comparing a predetermined number of the hypothesis words to a predetermined number of the artifact words, determining a match location in the text artifact where a specific number of the predetermined number of hypothesis words match a specific number of the predetermined number of artifact words, and altering the display on the display screen to display the match location on the display screen of the computer.

Abstract: A computer-implemented method for dynamically presenting a prewritten text in a graphical user interface is disclosed. The method comprises receiving a text artifact, storing the text artifact in a memory device of a computer, retrieving the text artifact, displaying the text artifact on the display screen of the computer, receiving a vocal input, generating a text file representing the words spoken in the vocal input, comparing a predetermined number of the hypothesis words to a predetermined number of the artifact words, determining a match location in the text artifact where a specific number of the predetermined number of hypothesis words match a specific number of the predetermined number of artifact words, and altering the display on the display screen to display the match location on the display screen of the computer.

Abstract: An arrangement and corresponding method are described for multi-domain natural language processing. Multiple parallel domain pipelines are used for processing a natural language input. Each domain pipeline represents a different specific subject domain of related concepts. Each domain pipeline includes a mention module that processes the natural language input using natural language understanding (NLU) to determine a corresponding list of mentions, and an interpretation generator that receives the list of mentions and produces a rank-ordered domain output set of sentence-level interpretation candidates. A global evidence ranker receives the domain output sets from the domain pipelines and produces an overall rank-ordered final output set of sentence-level interpretations.