Abstract: Training and/or utilizing a single neural network model to generate, at each of a plurality of assistant turns of a dialog session between a user and an automated assistant, a corresponding automated assistant natural language response and/or a corresponding automated assistant action. For example, at a given assistant turn of a dialog session, both a corresponding natural language response and a corresponding action can be generated jointly and based directly on output generated using the single neural network model. The corresponding response and/or corresponding action can be generated based on processing, using the neural network model, dialog history and a plurality of discrete resources. For example, the neural network model can be used to generate a response and/or action on a token-by-token basis.

Abstract: Training and/or utilizing a single neural network model to generate, at each of a plurality of assistant turns of a dialog session between a user and an automated assistant, a corresponding automated assistant natural language response and/or a corresponding automated assistant action. For example, at a given assistant turn of a dialog session, both a corresponding natural language response and a corresponding action can be generated jointly and based directly on output generated using the single neural network model. The corresponding response and/or corresponding action can be generated based on processing, using the neural network model, dialog history and a plurality of discrete resources. For example, the neural network model can be used to generate a response and/or action on a token-by-token basis.

Abstract: Training and/or utilizing a single neural network model to generate, at each of a plurality of assistant turns of a dialog session between a user and an automated assistant, a corresponding automated assistant natural language response and/or a corresponding automated assistant action. For example, at a given assistant turn of a dialog session, both a corresponding natural language response and a corresponding action can be generated jointly and based directly on output generated using the single neural network model. The corresponding response and/or corresponding action can be generated based on processing, using the neural network model, dialog history and a plurality of discrete resources. For example, the neural network model can be used to generate a response and/or action on a token-by-token basis.

Abstract: Disclosed is a method for managing bus channels between a physical layer and a media access channel layer of a network. The method includes determining a current rate of link speed on a media dependent interface side of a physical layer of a network architecture. When the current rate changes to yield a new link speed, the method includes determining a minimal (or near minimal) bus size required to implement the new link speed and switching to the minimal bus size between a physical coding sublayer of the physical layer and a reconciliation sublayer of a datalink layer of the network architecture. The method further includes switching to the minimal bus size between a physical coding sublayer of the physical layer and a reconciliation sublayer of a datalink layer of the network architecture.

Abstract: Methods, systems, and apparatus including computer programs encoded on a computer storage medium, for augmenting search engine index that indexes resources from a collection of resources.

Abstract: Disclosed is a method for managing bus channels between a physical layer and a media access channel layer of a network. The method includes determining a current rate of link speed on a media dependent interface side of a physical layer of a network architecture. When the current rate changes to yield a new link speed, the method includes determining a minimal (or near minimal) bus size required to implement the new link speed and switching to the minimal bus size between a physical coding sublayer of the physical layer and a reconciliation sublayer of a datalink layer of the network architecture. The method further includes switching to the minimal bus size between a physical coding sublayer of the physical layer and a reconciliation sublayer of a datalink layer of the network architecture.

Abstract: Methods, systems, and apparatus including computer programs encoded on a computer storage medium, for augmenting search engine index that indexes resources from a collection of resources.

Abstract: Techniques for managing information describing a hierarchy of relationships between educational institutions are described. According to various embodiments, first feature data describing a first school and second feature data describing a second school is accessed via one or more databases. A confidence score is then generated based on a machine learned model, the first feature data and the second feature data, the confidence score indicating a probability that the second school is a sub-school of the first school. Thereafter, based on a comparison of the confidence score to a threshold, is it is determined that the second school is a sub-school of the first school. Hierarchy information identifying a hierarchy of relationships between a plurality of schools is then generated or modified, the hierarchy information indicating that the second school is a sub-school of the first school.