Abstract: Systems and methods for adapting a neural network model on a hardware platform. An example method includes obtaining neural network model information comprising decision points associated with a neural network, with one or more first decision points being associated with a layout of the neural network. Platform information associated with a hardware platform for which the neural network model information is to be adapted is accessed. Constraints associated with adapting the neural network model information to the hardware platform are determined based on the platform information, with a first constraint being associated with a processing resource of the hardware platform and with a second constraint being associated with a performance metric. A candidate configuration for the neural network is generated via execution of a satisfiability solver based on the constraints, with the candidate configuration assigns values to the plurality of decision points.

Abstract: Systems and methods for adapting a neural network model on a hardware platform. An example method includes obtaining neural network model information comprising decision points associated with a neural network, with one or more first decision points being associated with a layout of the neural network. Platform information associated with a hardware platform for which the neural network model information is to be adapted is accessed. Constraints associated with adapting the neural network model information to the hardware platform are determined based on the platform information, with a first constraint being associated with a processing resource of the hardware platform and with a second constraint being associated with a performance metric. A candidate configuration for the neural network is generated via execution of a satisfiability solver based on the constraints, with the candidate configuration assigns values to the plurality of decision points.

Abstract: Systems and methods for adapting a neural network model on a hardware platform. An example method includes obtaining neural network model information comprising decision points associated with a neural network, with one or more first decision points being associated with a layout of the neural network. Platform information associated with a hardware platform for which the neural network model information is to be adapted is accessed. Constraints associated with adapting the neural network model information to the hardware platform are determined based on the platform information, with a first constraint being associated with a processing resource of the hardware platform and with a second constraint being associated with a performance metric. A candidate configuration for the neural network is generated via execution of a satisfiability solver based on the constraints, with the candidate configuration assigns values to the plurality of decision points.

Abstract: Systems and methods for adapting a neural network model on a hardware platform. An example method includes obtaining neural network model information comprising decision points associated with a neural network, with one or more first decision points being associated with a layout of the neural network. Platform information associated with a hardware platform for which the neural network model information is to be adapted is accessed. Constraints associated with adapting the neural network model information to the hardware platform are determined based on the platform information, with a first constraint being associated with a processing resource of the hardware platform and with a second constraint being associated with a performance metric. A candidate configuration for the neural network is generated via execution of a satisfiability solver based on the constraints, with the candidate configuration assigns values to the plurality of decision points.

Abstract: Systems and methods for adapting a neural network model on a hardware platform. An example method includes obtaining neural network model information comprising decision points associated with a neural network, with one or more first decision points being associated with a layout of the neural network. Platform information associated with a hardware platform for which the neural network model information is to be adapted is accessed. Constraints associated with adapting the neural network model information to the hardware platform are determined based on the platform information, with a first constraint being associated with a processing resource of the hardware platform and with a second constraint being associated with a performance metric. A candidate configuration for the neural network is generated via execution of a satisfiability solver based on the constraints, with the candidate configuration assigns values to the plurality of decision points.

Abstract: Displaying a visual representation of table, including generating a visual representation of a table including a plurality of dimensions, displaying the visual representation of the table, where a first dimension value corresponding to one of the plurality of dimensions is collapsed when the table is initially generated, and a second dimension value corresponding to one of the plurality of dimensions is expanded such that sub-dimension values for the first dimension value are not visible, and sub-dimension values for the second dimension value are visible, detecting that a user has placed a first dimension value manipulator corresponding to said first dimension value into an expand mode, and in response to detecting the first dimension value manipulator is in an expand mode, expanding the first dimension value in the displayed visual representation of the table such that the sub-dimension values for the first dimension value are visible.

Abstract: A method for applying policies to an email message includes receiving, by an inbound policy module in a protected network, message metadata of an email message. The method also includes determining, based on the message metadata, whether receiving the email message in the protected network is prohibited by at least one metadata policy. The method further includes blocking the email message from being forwarded to the protected network if receiving the email message in the protected network is prohibited by the metadata policy. In specific embodiments, the method includes requesting scan results data for the email message if receiving the email message in the protected network is not prohibited by one or more metadata policies. In further embodiments, the method includes receiving the scan results data and requesting the email message if receiving the email message in the protected network is not prohibited by one or more scan policies.

Abstract: A method for applying policies to an email message includes receiving, by an inbound policy module in a protected network, message metadata of an email message. The method also includes determining, based on the message metadata, whether receiving the email message in the protected network is prohibited by at least one metadata policy. The method further includes blocking the email message from being forwarded to the protected network if receiving the email message in the protected network is prohibited by the metadata policy. In specific embodiments, the method includes requesting scan results data for the email message if receiving the email message in the protected network is not prohibited by one or more metadata policies. In further embodiments, the method includes receiving the scan results data and requesting the email message if receiving the email message in the protected network is not prohibited by one or more scan policies.

Abstract: A method for applying policies to an email message includes receiving, by an inbound policy module in a protected network, message metadata of an email message. The method also includes determining, based on the message metadata, whether receiving the email message in the protected network is prohibited by at least one metadata policy. The method further includes blocking the email message from being forwarded to the protected network if receiving the email message in the protected network is prohibited by the metadata policy. In specific embodiments, the method includes requesting scan results data for the email message if receiving the email message in the protected network is not prohibited by one or more metadata policies. In further embodiments, the method includes receiving the scan results data and requesting the email message if receiving the email message in the protected network is not prohibited by one or more scan policies.

Abstract: A method for applying policies to an email message includes receiving, by an inbound policy module in a protected network, message metadata of an email message. The method also includes determining, based on the message metadata, whether receiving the email message in the protected network is prohibited by at least one metadata policy. The method further includes blocking the email message from being forwarded to the protected network if receiving the email message in the protected network is prohibited by the metadata policy. In specific embodiments, the method includes requesting scan results data for the email message if receiving the email message in the protected network is not prohibited by one or more metadata policies. In further embodiments, the method includes receiving the scan results data and requesting the email message if receiving the email message in the protected network is not prohibited by one or more scan policies.

Abstract: Displaying a visual representation of table, including generating a visual representation of a table including a plurality of dimensions, displaying the visual representation of the table, where a first dimension value corresponding to one of the plurality of dimensions is collapsed when the table is initially generated, and a second dimension value corresponding to one of the plurality of dimensions is expanded such that sub-dimension values for the first dimension value are not visible, and sub-dimension values for the second dimension value are visible, detecting that a user has placed a first dimension value manipulator corresponding to said first dimension value into an expand mode, and in response to detecting the first dimension value manipulator is in an expand mode, expanding the first dimension value in the displayed visual representation of the table such that the sub-dimension values for the first dimension value are visible.

Abstract: A method for applying policies to an email message includes receiving, by an inbound policy module in a protected network, message metadata of an email message. The method also includes determining, based on the message metadata, whether receiving the email message in the protected network is prohibited by at least one metadata policy. The method further includes blocking the email message from being forwarded to the protected network if receiving the email message in the protected network is prohibited by the metadata policy. In specific embodiments, the method includes requesting scan results data for the email message if receiving the email message in the protected network is not prohibited by one or more metadata policies. In further embodiments, the method includes receiving the scan results data and requesting the email message if receiving the email message in the protected network is not prohibited by one or more scan policies.

Abstract: A method for applying policies to an email message includes receiving, by an inbound policy module in a protected network, message metadata of an email message. The method also includes determining, based on the message metadata, whether receiving the email message in the protected network is prohibited by at least one metadata policy. The method further includes blocking the email message from being forwarded to the protected network if receiving the email message in the protected network is prohibited by the metadata policy. In specific embodiments, the method includes requesting scan results data for the email message if receiving the email message in the protected network is not prohibited by one or more metadata policies. In further embodiments, the method includes receiving the scan results data and requesting the email message if receiving the email message in the protected network is not prohibited by one or more scan policies.

Abstract: A hardware fixture for doors, cabinets, drawers or other fixtures, has an interchangeable decorative or informational attachment. The hardware fixture includes a knob or pull body receiving the attachment, which can be shaped as a disc. The surface of the body may be recessed or surrounded by a rim that protects an edge of the decorative member against inadvertent dislodgment. The body includes a magnet or ferrous portion and the interchangeable decorative attachment similarly includes a magnet or ferrous portion that is attracted to the body. The magnetic surfaces abut only part way across the central portion of the knob body. The decorative attachment can be dislodged when desired by tipping it. Downward pressure on the decorative member over a gap with the knob lifts the opposite edge of the decorative attachment such that it can be grasped.

Abstract: The invention relates to an electrical connector that can be used to electrically connect a daughtercard to a backplane.

Abstract: The present invention is directed to a high density electrical connector which can provide 80 or more twinax connections per linear inch in a 20 millimeter card slot. In a typical electronic system package, 20 millimeters is the spacing from center line to center line of the adjacent parallel daughtercards. Twinax cable is coaxial cable that contains two .inner conducting wires rather than one. The two inner conducting wires provide two physical channels. Coaxial cable is called “coaxial” because it includes one physical channel that carries the signal surrounded (after a layer of insulation) by another concentric physical channel, both running along the same axis. The outer channel serves as ground.