Abstract: An apparatus to facilitate execution of non-linear functions operations is disclosed. The apparatus comprises accelerator circuitry including a compute grid having a plurality of processing elements to execute neural network computations, store values resulting from the neural network computations, and perform piecewise linear (PWL) approximations of one or more non-linear functions using the stored values as input data.

Abstract: Hardware accelerators for accelerated grouped convolution operations. A first buffer of a hardware accelerator may receive a first row of an input feature map (IFM) from a memory. A first group comprising a plurality of tiles may receive a first row of the IFM. A plurality of processing elements of the first group may compute a portion of a first row of an output feature map (OFM) based on the first row of the IFM and a kernel. A second buffer of the accelerator may receive a third row of the IFM from the memory. A second group comprising a plurality of tiles may receive the third row of the IFM. A plurality of processing elements of the second group may compute a portion of a third row of the OFM based on the third row of the IFM and the kernel as part of a grouped convolution operation.

Abstract: An apparatus to facilitate execution of non-linear functions operations is disclosed. The apparatus comprises accelerator circuitry including a compute grid having a plurality of processing elements to execute neural network computations, store values resulting from the neural network computations, and perform piecewise linear (PWL) approximations of one or more non-linear functions using the stored values as input data.

Abstract: An apparatus to facilitate acceleration of machine learning operations is disclosed. The apparatus comprises accelerator circuitry including a first set of processing elements to perform first computations including matrix multiplication operations, a second set of processing elements to perform second computations including sum of elements of weights and offset multiply operations and a third set of processing elements to perform third computations including sum of elements of inputs and offset multiply operations, wherein the second and third computations are performed in parallel with the first computations.

Abstract: Hardware accelerators for accelerated grouped convolution operations. A first buffer of a hardware accelerator may receive a first row of an input feature map (IFM) from a memory. A first group comprising a plurality of tiles may receive a first row of the IFM. A plurality of processing elements of the first group may compute a portion of a first row of an output feature map (OFM) based on the first row of the IFM and a kernel. A second buffer of the accelerator may receive a third row of the IFM from the memory. A second group comprising a plurality of tiles may receive the third row of the IFM. A plurality of processing elements of the second group may compute a portion of a third row of the OFM based on the third row of the IFM and the kernel as part of a grouped convolution operation.

Abstract: An apparatus to facilitate acceleration of machine learning operations is disclosed. The apparatus comprises accelerator circuitry including a first set of processing elements to perform first computations including matrix multiplication operations, a second set of processing elements to perform second computations including sum of elements of weights and offset multiply operations and a third set of processing elements to perform third computations including sum of elements of inputs and offset multiply operations, wherein the second and third computations are performed in parallel with the first computations.

Abstract: A method and system for optimizing network access control are disclosed. For example, the method includes receiving an access request to a network from a device. Then, the method determines if each one of one or more critical updates of said device is current and if each one of one or more non-critical updates of the device is current. The method concludes by granting the access request to the network if each one of the one or more critical updates of the device is current, even if at least one non-critical update of the one or more non-critical updates of the device is not current.

Abstract: A virtual file management system provides user access to managed content on mobile devices. The system comprises storage domains storing the managed content distributively using file systems, and a data infrastructure that organizes the managed content into a virtual file system that maintains information of storage domain specific file system primitives for accessing corresponding portions of the managed content. The data infrastructure, which maintains metadata of the storage domains and the mobile devices, comprises a policy definition and decision component that maintains policies defining controls for permissible operations on the managed content, the permissible operations including the file system primitives.

Abstract: A virtual file management system provides user access to managed content on mobile devices. The system comprises storage domains storing the managed content distributively using file systems, and a data infrastructure that organizes the managed content into a virtual file system that maintains information of storage domain specific file system primitives for accessing corresponding portions of the managed content. The data infrastructure, which maintains metadata of the storage domains and the mobile devices, comprises a policy definition and decision component that maintains policies defining controls for permissible operations on the managed content, the permissible operations including the file system primitives.

Abstract: A method, apparatus, and electronic device for conforming integrity of a client device 106 are disclosed. A memory 1100 may store a policy tag 404 associated with a subgroup of a group of policies 1102 and having a tag timestamp. A network interface 1060 may receive the certificate of health 300 from the client device 106. A processor 1010 may extract from the certificate of health a certificate timestamp 302 and a policy tag 304. The processor 1010 may access the tag timestamp. The processor 1010 may execute a comparison of the certificate timestamp 302 with the tag timestamp. The network interface 1060 may grant access to a network 104 based in part upon the comparison.

Abstract: A method and system for optimizing network access control are disclosed. For example, the method includes receiving an access request to a network from a device. Then, the method determines if each one of one or more critical updates of said device is current and if each one of one or more non-critical updates of the device is current. The method concludes by granting the access request to the network if each one of the one or more critical updates of the device is current, even if at least one non-critical update of the one or more non-critical updates of the device is not current.

Abstract: A mobile multimedia framework API is provided that is capable of operation in mobile hardware devices. The mobile multimedia framework API includes a playback interface, controls, an event listener, an error list, exceptions, events, control types, and protocol handlers. Advantageously, a memory size of the mobile multimedia framework API is less than 100 kilobytes.

Abstract: A method and system is provided for customizing Java API implementations. A plurality of class files is obtained, wherein the class files are capable of being used together to provide a plurality of functional behaviors to an application. A feature marker is then associated with each of the class files based on a functionality provided by the class file. Each feature marker indicates a particular functionality provided by the plurality of class files. Particular class files are then selected from the plurality of class files based on the feature markers associated with the particular class files. The plurality of class files can comprise a first JAR file, and the selected class files can be used to generate a second JAR file. The second JAR file can also exclude class files that are not selected. Further, non-Java based native files, if any, utilized by the plurality of class files can also be customized to a subset of the original native files in a similar manner.

Abstract: An invention is provided for creating a mobile multimedia framework API capable of operation in mobile hardware devices is disclosed. All API component access parameters are set to utilize a synchronous programming model. Then, the API components are set to use a pull paradigm, and all master/slave functionality is removed. The resulting mobile multimedia framework API has a memory size of less than 100 kilobytes.

Abstract: An invention is provided for customizing Java API implementations. A plurality of class files is obtained, wherein the class files are capable of being used together to provide a plurality of functional behaviors to an application. A feature marker is then associated with each of the class files based on a functionality provided by the class file. Each feature marker indicates a particular functionality provided by the plurality of class files. Particular class files are then selected from the plurality of class files based on the feature markers associated with the particular class files. In one aspect, the plurality of class files can comprise a first JAR file, and the selected class files can be used to generate a second JAR file. The second JAR file can also exclude class files that are not selected. Further, non-Java based native files, if any, utilized by the plurality of class files can also be customized to a subset of the original native files in a similar manner.

Abstract: A mobile multimedia framework API is provided that is capable of operation in mobile hardware devices. The mobile multimedia framework API includes a playback interface, controls, an event listener, an error list, exceptions, events, control types, and protocol handlers. Advantageously, a memory size of the mobile multimedia framework API is less than 100 kilobytes.

Abstract: An invention is provided for creating a mobile multimedia framework API capable of operation in mobile hardware devices is disclosed. All API component access parameters are set to utilize a synchronous programming model. Then, the API components are set to use a pull paradigm, and all master/slave functionality is removed. The resulting mobile multimedia framework API has a memory size of less than 100 kilobytes.