Abstract: Technologies are disclosed for to utilizing a virtual assistant to as a meeting agenda facilitator. The virtual assistant may start the meeting, control the flow of the meeting, introduce agenda items, receive comments, display meeting materials associated with the meeting, generate/keep track of follow-up actions, create a meeting record, and the like. In some examples, virtual assistant accesses a meeting agenda to determine what agenda items are to be discussed, what content is to be presented, as well as summaries of the meeting and the agenda items. The meeting agenda may also include an expected duration of each of the agenda items that may be used to keep the meeting on time. The virtual assistant may also generate follow-up actions to be performed. Still yet, the virtual assistant may create a meeting record that can include information recorded from the meeting and is associated with the agenda items discussed.

Abstract: A method, apparatus, and computer program for controlling allocation of control message fields in uplink transmission in a cellular telecommunication system are presented. Uplink control message fields are allocated to the resources of a physical uplink shared traffic channel according to an uplink transmission scheme selected for a user terminal. The control message fields are allocated so that transmission performance of the control messages is optimized for the selected uplink transmission scheme.

Abstract: Some embodiments include a semiconductor device having a stack structure including a source comprising polysilicon, an etch stop of oxide on the source, a select gate source on the etch stop, a charge storage structure over the select gate source, and a select gate drain over the charge storage structure. The semiconductor device may further include an opening extending vertically into the stack structure to a level adjacent to the source. A channel comprising polysilicon may be formed on a side surface and a bottom surface of the opening. The channel may contact the source at a lower portion of the opening, and may be laterally separated from the charge storage structure by a tunnel oxide. A width of the channel adjacent to the select gate source is greater than a width of the channel adjacent to the select gate drain.

Abstract: In one embodiment, a computer networking device calculates a first hash value for an identifier of a group of computing devices, as well as a second hash value for the identifier of the group of computing devices, with each hash value being at least in part on the identifier of the group of computing devices and an identifier of the respective interface. The computer networking device may also analyze the first hash value with respect to the second hash value and select the first interface for association with the identifier of the group of computing devices based at in part on the analyzing. The computer networking device may further store an indication that the identifier of the group of computing devices is associated with the first interface.

Abstract: Methods, systems, and computer-readable media for batch processing with random access for transaction history are disclosed. A batch processing system receives a batch comprising records of events, including a first record of a first event and a second record of a second event. The system assigns the first and second records to a group based (at least in part) on determining that the events are related. The system determines that the group is related to a match set comprising one or more prior events. The system updates one or more values in the match set based (at least in part) on the first and second records. The system stores the updated match set and one or more additional match sets using a storage object. The system retrieves the match set and not the one or more additional match sets from the storage object using an index.

Abstract: Systems and methods that facilitate multilevel hysteresis voltage control methods for cascaded multilevel voltage modulators having a plurality of power cells connected in series and has any positive integer number of output voltage levels to control any unipolar voltage on the load of the voltage modulator, and transfer electrical power from an electrical grid via AC/DC converters or directly from energy storage elements of the power cells to that load. A method of operational rotation of the power cells of a multilevel voltage modulator, which ensures an equal power sharing among the power cells and voltage balancing of the energy storage elements of the power cells of the modulator.

Abstract: Image sensors include a photodiode disposed in a semiconductor substrate and a transistor operatively coupled to the photodiode. At least three substrate trench structures are formed in the semiconductor substrate, defining two nonplanar structures, each having a plurality of sidewall portions. An isolation layer includes at least three isolation layer trench structures, each being disposed in a respective one of the three substrate trench structures. A gate includes three fingers, each being disposed in a respective one of the three isolation layer trench structures. An electron channel of the transistor extends along the plurality of sidewall portions of the two nonplanar structures in a channel width plane.

Abstract: Systems, methods, apparatus, computer program code and means are provided for bulk costing using an application server which receives a bulk costing request from a remote user device, the bulk costing request including a list of items to be costed and a set of input variables. A final list of components and assemblies is generated including a scenario for each item to be costed for each combination of input variables. A deep costing analysis is performed of the final list, resulting in an estimated cost for each of the items to be costed for each of the input variables.

Abstract: Devices and techniques are generally described for articulated three-dimensional pose tracking. In some examples, a plurality of frames of image data captured by one or more cameras may be received. First feature data representing the plurality of frames of image data may be determined using a backbone network. The first feature data may be projected into three-dimensional (3D) space. In some examples, 3D location data describing respective 3D locations of one or more persons represented by the first feature data projected in the 3D space may be determined. The first feature data and the 3D location data may be sent to a four-dimensional (4D) convolutional neural network (CNN). The 4D CNN may generate second feature data comprising respective 3D representations of the one or more persons. Three dimensional pose data representing articulated 3D pose information for the one or more persons may be generated.

Abstract: Aspects of the present disclosure relate to systems and methods for receiving, managing, and displaying annotations on documents in real-time. A user (e.g., an author of a document) uploads a document into a real-time annotation system, which may then generate a composite presentation based on the uploaded document. The composite presentation includes all the content of the document presented in a specially configured graphical user interface to receive and manage annotations from a plurality of user devices.

Abstract: Techniques for a device with NUMA a memory architecture to migrate virtual resources between NUMA nodes to reduce resource contention between virtual resources running on the NUMA nodes. In some examples, the device monitors various metrics and/or operations of the NUMA nodes and/or virtual resources, and detect events that indicate that virtual resources running on a same NUMA node are contending, or are likely to contend, for computing resources of the NUMA node. Upon detecting such an event, the device may migrate a virtual resource from the NUMA node to another NUMA node on the device that has an availability of computing resources. The device may then migrate the virtual resource from the overcommitted NUMA node onto the NUMA node that has availability to run the virtual resource. In this way, devices may reduce resource contention among virtual resources running on a same NUMA node.

Abstract: To improve utilization of a systolic array, each row of the array is provided with a number of general purpose row input data buses. Each of the general purpose row input data buses can be operable to transfer either feature map (FMAP) input elements or weight values into the processing elements of the corresponding row of the array. By using such general purpose row input data buses, concurrent matrix multiplications as well as faster background weight loading can be achieved in the array.

Abstract: The technique herein optimizes delivery performance from a content delivery network (CDN) edge node, preferably by using knowledge of radio link allocation behavior. According to the technique herein, an IP address of a cellular end user (or equivalent) is recognized by the CDN edge. Then, preferably by tracking idle times both in consecutive downloading bursts in the end user session and in an on/off style radio link allocation pattern, the CDN edge adjusts transport layer network protocol parameters. The approach (which leverages the cross-layer intelligence obtainable from existing CDN infrastructure) thus enables the transport layer protocol implemented at the edge to optimize the delivery performance over the cellular mobile network.

Abstract: A memory device and a forming method thereof are provided. The memory device includes: a semiconductor substrate, wherein multiple active regions are formed in the semiconductor substrate, and the multiple active regions are separated by multiple first trenches extending along a first direction and multiple second trenches extending along a second direction; a third trench, extending along the first direction and located in the semiconductor substrate at the bottom of the first trench; a bit line doped region, located in the semiconductor substrate on two sides of the third trench; a gate dielectric layer, located on a sidewall surface of the first trench and a sidewall surface of the second trench; a first dielectric layer that fills the third trench; a metal gate, located in the second trench and the first trench on the first dielectric layer.

Abstract: Systems and methods for a wireless data protocol are disclosed. For example, a packet fragmentation scheme is determined and utilized to fragment high bit rate packets. A packet spacing scheme and a number of parity slots to utilize is determined, and then packet fragment representations are generated utilizing the packet spacing scheme and the number of parity slots. The packet fragment representations are appended to their corresponding packet fragments and sent to a receiving device. The receiving device may then utilize the packet fragment representations to reconstruct packets that are lost during transmission.

Abstract: Systems, apparatuses, and methods related to an extended memory neuromorphic component for performing operations in memory are described. An example apparatus can include a plurality of computing devices. Each of the computing devices can include a processing unit and a memory array. The example apparatus can further include a communication subsystem coupled to the at least one of the plurality of computing devices and to a neuromorphic component. At least one of the plurality of computing devices can receive a request from a host to perform an operation, receive an indication of data to be access in a memory device to perform the operation, and send an indication to the neuromorphic component to monitor the data to be accessed. The neuromorphic component can intercept data and determine that a portion of the data should be flagged.

Abstract: A computer-implemented method for reencrypting data. A key management service receives a web service application programming interface or other request to reencrypt data from a first key to a second key, where the first key and the second key are managed by the key management service on behalf of a user of the service. The key management service response to the request by performing the associated operations and providing a response with the reencrypted data.

Abstract: Iterative signal processing. At communication hardware, a signal is received from a transmission medium. The signal has characteristics that obscure data or a signal of interest in the signal. The signal is processed at a first signal processor, which is an iterative processor that performs signal processing in cycles whereby successive cycles: improve the performance of processing of the processor itself over previous cycles, or improve the output from the processor. The signal is processed at one or more second signal processors. Extrinsic data, with respect to the first signal processor is produced as a result. The extrinsic data is provided to the first signal processor and used to counter the effects of the data or signal of interest being obscured in the signal, while the first signal processor is intracycle of a first processing cycle.

Abstract: Systems, methods, and computer-readable media are disclosed for language-agnostic subtitle drift detection and correction. A method may include determining subtitles and/or captions from media content (e.g., videos), the subtitles and/or captions corresponding to dialog in the media content. The subtitles may be broken up into segments which may be analyzed to determine a likelihood of drift (e.g., a likelihood that the subtitles are out of synchronization with the dialog in the media content) for each segment. For segments with a high likelihood of drift, the subtitles may be incrementally adjusted to determine an adjustment that eliminates and/or reduces the amount of drift and the drift in the segment may be corrected based on the drift amount detected. A linear regression model and/or human blocks determined by human operators may be used to otherwise optimize drift correction.

Abstract: Technologies are provided for reporting engaged impressions of directed content. Some embodiments include a computing device that can initiate a screensaver session. The computing device also can cause presentation of a directed content asset on a display device functionally coupled to the computing device. The computing device can then update a queue to add a record of an impression of the directed content asset, where the queue is retained in a memory device of the computing device. The computing device can receive a signal indicative of user activity. The computing device can send, to an attribution server device, first data identifying one or more directed content assets corresponding to one or more records of the queue.