Abstract: One embodiment sets forth a technique for generating a tree structure within a computer memory for storing sparse data. The technique includes dividing a matrix into a first plurality of equally sized regions. The technique also includes dividing at least one region in the first plurality of regions into a second plurality of regions, where the second plurality of regions includes a first region and one or more second regions that have a substantially equal number of nonzero matrix values and are formed within the first region. The technique further includes creating the tree structure within the computer memory by generating a first plurality of nodes representing the first plurality of regions, generating a second plurality of nodes representing the second plurality of regions, and grouping, under a first node representing the first region, one or more second nodes representing the one or more second regions.

Abstract: A myringotomy device includes a housing; an elongated tube extending from the housing; and a retractable cutting tool extendable through the elongated tube, the cutting tool comprising a blade. The cutting tool is configured such that when advanced, the blade of the cutting tool extends beyond a distal end of the elongated tube. The cutting tool is also configured such that when retracted, the blade is retracted into the elongated tube and a fluid conduit is created from the distal end of the elongated tube to the housing.

Abstract: Systems and methods electrically connect a sealed speaker and speaker back volume without wires. An electronic device includes a speaker with a first and second speaker contacts, a speaker box coupled with the speaker, and a printed circuit board (PCB) having a first and second contact pads. A speaker gasket is positioned between the speaker and the speaker box to form a seal. First and second spring strips electrically connect the first and second contact pads to the first and second speaker contacts, respectively, and each has a flat, thin cross-section in at least a portion extending between the speaker gasket and the speaker box. The first and second spring strips do not adversely affect the seal, do not require glue or sealer, and electrically connect with the first and second contact pads as the enclosure of the electronic device is closed.

Abstract: A dark fiber dense wavelength division multiplexing service path design microservice (ddSPDmS) can provide a scalable self-contained meta-data driven approach for a flexible implementation of a dark fiber dense wavelength division multiplexing (DWDM) service path design solution. The service plan design solution can be used as a standalone solution or integrated with a network management application. In order to manage a large volume of circuit designs, multiple microservices can accept application program interface (API) requests in a cloud environment. Permission can then be given to any application to use the API to make a call to the design and inventory. Additionally, metadata templates can be designed to support a node, a link, and/or a topology for the microservices.

Abstract: Embodiments include methods, systems and non-transitory computer-readable computer readable media including instructions for executing a prefetch kernel that includes memory accesses for prefetching data for a processing kernel into a memory, and, subsequent to executing at least a portion of the prefetch kernel, executing the processing kernel where the processing kernel includes accesses to data that is stored into the memory resulting from execution of the prefetch kernel.

Abstract: PAM encoding techniques that leverage unused idle periods in channels between data transmissions to apply longer but more energy-efficient codes. To improve energy savings, multiple sparse encoding schemes may be utilized selectively to fit different sized gaps in the traffic. These approaches may provide energy reductions, for example with memory READ and WRITE traffic, when transferring 4-bit data using 3-symbol sequences.

Abstract: A myringotomy device includes a housing; an elongated tube extending from the housing; and a retractable cutting tool extendable through the elongated tube, the cutting tool comprising a blade. The cutting tool is configured such that when advanced, the blade of the cutting tool extends beyond a distal end of the elongated tube. The cutting tool is also configured such that when retracted, the blade is retracted into the elongated tube and a fluid conduit is created from the distal end of the elongated tube to the housing.

Abstract: Embodiments of the disclosed subject matter provide a device that includes an organic light emitting device (OLED), and a drive circuit to control the operation of the OLED, comprising a response time accelerator thin film transistor (TFT) configured to short or reverse bias the OLED for a predetermined period of time during a frame time. Other embodiments include an OLED having a plurality of sub-pixels, where one or more of sub-pixels configured to emit light of at least a first color comprises a first emissive area and a second emissive area that are independently controllable, where the first emissive area is larger than the second emissive area. The controller is configured to control the second emissive area to have (i) a higher brightness, and/or (ii) a higher current density than the first emissive area for a first sub-pixel luminance level that is less than a maximum luminance.

Abstract: An aspect of the invention provides a method of performing a distributed task over a network comprising a plurality of nodes. The method comprises: a plurality of network nodes observing (300) data; applying a first linear code function to the data observed by at least one network node of the plurality of network nodes to obtain (302) at least one function output; applying errors (304) to the at least one function output; a query node selected from the network nodes performing (308) a mixing procedure to aggregate node observations to obtain a first set of aggregated values until a stopping criteria (306) is satisfied; applying (312) a second linear code function to the set of aggregated values to obtain a second set of aggregated values returned to their observed domain; and the query node outputting (314) the second set of aggregated values.

Abstract: A dark fiber dense wavelength division multiplexing service path design microservice (ddSPDmS) can provide a scalable self-contained meta-data driven approach for a flexible implementation of a dark fiber dense wavelength division multiplexing (DWDM) service path design solution. The service plan design solution can be used as a standalone solution or integrated with a network management application. In order to manage a large volume of circuit designs, multiple microservices can accept application program interface (API) requests in a cloud environment. Permission can then be given to any application to use the API to make a call to the design and inventory. Additionally, metadata templates can be designed to support a node, a link, and/or a topology for the microservices.

Abstract: One embodiment sets forth a technique for generating a tree structure within a computer memory for storing sparse data. The technique includes dividing a matrix into a first plurality of equally sized regions. The technique also includes dividing at least one region in the first plurality of regions into a second plurality of regions, where the second plurality of regions includes a first region and one or more second regions that have a substantially equal number of nonzero matrix values and are formed within the first region. The technique further includes creating the tree structure within the computer memory by generating a first plurality of nodes representing the first plurality of regions, generating a second plurality of nodes representing the second plurality of regions, and grouping, under a first node representing the first region, one or more second nodes representing the one or more second regions.

Abstract: One embodiment sets forth a technique for performing one or more matrix multiplication operations based on a first matrix and a second matrix. The technique includes receiving data associated with the first matrix from a first traversal engine that accesses nonzero elements included in the first matrix via a first tree structure. The technique also includes performing one or more computations on the data associated with the first matrix and the data associated with the second matrix to produce a plurality of partial results. The technique further includes combining the plurality of partial results into one or more intermediate results and storing the one or more intermediate results in a first buffer memory.

Abstract: One embodiment sets forth a technique for performing matrix operations. The technique includes traversing a tree structure to access one or more non-empty regions within a matrix. The tree structure includes a first plurality of nodes and a second plurality of nodes corresponding to non-empty regions in the matrix. The first plurality of nodes includes a first node representing a first region and one or more second nodes that are children of the first node and represent second region(s) with an equal size formed within the first region. The second plurality of nodes include a third node representing a third region and one or more fourth nodes that are children of the third node and represent fourth region(s) with substantially equal numbers of non-zero matrix values formed within the third region. The technique also includes performing matrix operation(s) based on the non-empty region(s) to generate a matrix operation result.

Abstract: Embodiments include methods, systems and non-transitory computer-readable computer readable media including instructions for executing a prefetch kernel with reduced intermediate state storage resource requirements. These include executing a prefetch kernel on a graphics processing unit (GPU), such that the prefetch kernel begins executing before a processing kernel. The prefetch kernel performs memory operations that are based upon at least a subset of memory operations in the processing kernel.

Abstract: Implementations include receiving two or more time-series data sequences representative of a target process executed within a physical environment, executing automated time-series process segmentation to provide a plurality of subsequence segments for each of the two or more time-series data sequences, each subsequence segment corresponding to a phase of the target process, processing the two or more subsequence segments using at least one time-series transformation to provide a feature data set for each subsequence segment, applying each feature data set to provide time-series models for anomaly detection and forecasting, respectively, each time-series model being provided as one of a recurrent neural network (RNN), a convolution neural network (CNN), and a generative adversarial network (GAN), determining anomaly scores based on the time-series models, and selectively providing an alert to one or more users, each alert indicating at least one anomaly and a respective probability.

Abstract: A combined on-package and off-package memory system uses a custom base-layer within which are fabricated one or more dedicated interfaces to off-package memories. An on-package processor and on-package memories are also directly coupled to the custom base-layer. The custom base-layer includes memory management logic between the processor and memories (both off and on package) to steer requests. The memories are exposed as a combined memory space having greater bandwidth and capacity compared with either the off-package memories or the on-package memories alone. The memory management logic services requests while maintaining quality of service (QoS) to satisfy bandwidth requirements for each allocation. An allocation may include any combination of the on and/or off package memories. The memory management logic also manages data migration between the on and off package memories.

Abstract: A balun is enhanced with design features that extend the operational bandwidth of the balun allowing the balun to operate at lower frequencies. The design enhancements also suppress resonances that otherwise cause sudden power drops at a resonance frequency while a load is connected between the balun's differential outputs.