Abstract: Systems and techniques for attention calculation optimization are described. An attention calculation system identifies tensor dimensions based on a characteristic of a tensor multiplication engine. In some examples, the tensor dimensions are matrix dimensions, for instance if the characteristic indicates that the tensor multiplication engine is optimized for matrix multiplication. The attention calculation system groups at least a subset of query data into at least one query tensor having the tensor dimensions. The attention calculation system groups at least a subset of key data into at least one key tensor having the tensor dimensions. The attention calculation system determines, using the tensor multiplication engine, a tensor multiplication including the at least one query tensor and the at least one key tensor to generate output data.

Abstract: Meeting notes are generated within a communication session. In one embodiment, the system connects to a communication session with a plurality of participants; receives, during the communication session, updates to a transcript of a conversation between the participants produced during the communication session, the transcript including timestamps for a plurality of utterances associated with speaking participants; receives a request from one of the participants to generate a meeting note which includes a subset of utterances from the transcript; determines a starting timestamp and an ending timestamp for utterances representing the boundaries of the meeting note; generates, in real time during the communication session, a meeting note where the utterances for the starting timestamp and ending timestamp are the starting utterance and ending utterance of the meeting note, respectively; and presents to at least the participant, within a user interface, the generated meeting note.

Abstract: The technology relates to systems and methods for generating context-aware, dynamic visualizations. In an example, a method includes extracting a dynamic context signal; based on the dynamic context signal, retrieving a first visualization; causing a display of the first visualization as part of an application user interface; at an expiration of a refresh period, extracting an updated dynamic context signal; based on the updated dynamic context signal, retrieving a second visualization; and replacing the first visualization with the second visualization as part of an application user interface.

Abstract: The present invention generally relates to a virtual reality (VR) navigation system using artificial intelligence (AI), thus enabling providing of individualized and context-aware navigation experience in VR environments. It consists of an AI-inclusive central processing unit (CPU) that receives and processes real-time data from multiple hardware sensors deployed in a wearable VR headset, e.g., inertial measurement units (IMUs), eye-tracking sensors, biometric ones. This is AI-driven CPU that dynamically changes the navigation path and sensory outputs such as visual, audio, haptic based on user input/interaction & environmental context. It consists of a hand-held controller with numerous input interfaces and a wireless communication module, enabling it to be easily linked or integrated into several VR applications.

Abstract: In described examples, an integrated circuit (IC) includes an isolation, an input/output (IO), and a low power mode (LPM) control logic. The isolation includes a level shift with pull-down configured to weakly pull down the voltage of signals that travel through the isolation. The IO includes an input and a physical connector for coupling to a power management IC. The IO provides an asserted-low LPM enable signal to the physical connector in response to the IO input. An output of the LPM control logic is coupled via the isolation to the input of the IO. The LPM control logic provides a high voltage signal to the input of the IO as a default during power on reset (POR) of the IC. The pull-down pulls the LPM enable signal voltage to the asserted low voltage in response to a voltage of the LPM enable signal falling below a threshold.

Abstract: Systems and methods for improving the power, performance, and area (PPA) for a voltage mode driver and on die termination (ODT). A voltage mode driver having first and second circuits in a pulldown design. The first circuit has a plurality of nMOS devices in parallel, the plurality of nMOS devices being common to a first resistor. The second circuit is in parallel with the first circuit and has an nMOS device in series with a second resistor. The second circuit is configured to be enabled when the pulldown impedance of the first circuit, with the second circuit disabled and all of the nMOS devices of the first circuit turned on, is greater than a desired pulldown impedance. The voltage mode driver may also be a pullup design, or have both pulldown and pullup stages.

Abstract: Methods and systems provide for real-time meeting notes within a communication session.

Abstract: An apparatus for is disclosed for precoding radar receive signals. The apparatus includes a wireless transceiver configured to transmit a radar transmit signal using at least one antenna. The radar transmit signal is modulated in frequency across multiple cycles based on one or more frequency-modulation schemes. The wireless transceiver is also configured to receive at least two radar receive signals using at least two other antennas. The at least two radar receive signals include versions of the radar transmit signal that are reflected by at least one object. Each of the at least two radar receive signals is modulated in frequency across the multiple cycles based on the one or more frequency-modulation schemes of the radar transmit signal. The wireless transceiver is further configured to precode the at least two radar receive signals across the multiple cycles to generate precoded radar receive signals that are orthogonal to each other.

Abstract: Fiducial patterns that produce diffraction patterns at a camera sensor are etched or otherwise provided on the surface of a cover glass (CG) in front of a camera. An active light source injects light into the cover glass or into a diffractive optical element to strengthen the signal from the fiducial pattern, or alternatively an active light source is used to reflect light off a reflective fiducial pattern on the cover glass, thus requiring fewer frames to capture and process the diffraction pattern caused by the fiducial pattern.

Abstract: A method for preparing a credential package includes providing access to a credential record of a plurality of credential records stored in a database system. The credential record includes information identifying a credential candidate and credential information associated with the credential candidate. The method further comprising receiving a credential document associated with the credential information, receiving credential document information associated with the credential document, and storing the credential document in a distributed ledger system comprising a plurality of nodes.

Abstract: A vehicle charging apparatus is described herein, which may include a battery pack comprising a plurality of individual batteries, a power input port receiving electrical power at a first wattage, an AC-to-DC conversion circuit configured to provide DC power to charge groups of batteries in the plurality of individual batteries, a power conversion circuit configured to condition a DC output of at least one group of batteries to provide a charging current output to a vehicle via a coupling, and a processing circuit configured to control the power conversion circuit to provide the charging current at a second wattage greater than the first wattage. The first wattage may be actively or inherently limited to a level less than the second wattage in order to provide fast DC charging with a limited power input.

Abstract: An adjustable insulating device is provided. An example device includes a housing and an extendable member at least partially located within the housing. The extendable member moves between a retracted position and an extended position. The device further includes an adjustment mechanism operably coupled with the extendable member. The adjustment mechanism causes translation of the extendable member between the retracted position and the extended position. The adjustment mechanism includes a first roller rotationally engaged with the extendable member at a first position that houses a first motor, and a second roller rotationally engaged with the extendable member at a second position that houses a second motor. An output of the first motor and an output of the second motor are configured to cause translation of the extendable member between the retracted position and the extended position, via the first roller and the second roller.

Abstract: Domain-specific parameters may be used for tuning speech processing. A pre-trained transformer-based language model may train domain-specific parameters using domain-specific unlabeled text data. This domain-specific parameters can then be appended to candidate texts produced by a speech model on received speech data and input to the transformer-based language model to score the candidate texts. The scores of the candidate texts determined using the pre-trained transformer-based language model can then be used to select a candidate text for further speech processing.

Abstract: In described examples, an integrated circuit (IC) includes an isolation, an input/output (IO), and a low power mode (LPM) control logic. The isolation includes a level shift with pull-down configured to weakly pull down the voltage of signals that travel through the isolation. The IO includes an input and a physical connector for coupling to a power management IC. The IO provides an asserted-low LPM enable signal to the physical connector in response to the IO input. An output of the LPM control logic is coupled via the isolation to the input of the IO. The LPM control logic provides a high voltage signal to the input of the IO as a default during power on reset (POR) of the IC. The pull-down pulls the LPM enable signal voltage to the asserted low voltage in response to a voltage of the LPM enable signal falling below a threshold.

Abstract: Disclosed is a system for providing machine learning aided diagnostics and prognostics for large distributed systems. A diagnostics module applies two-tiered analysis to detect anomalous behavior of the large scale distributed system. First, multivariate telemetry and event data emitted from the large scale distributed systems is collected by a diagnostics component, which applies multivariate analysis to identify of set of N-anomalies. Second, univariate telemetry and event data is obtained by the diagnostics component, which applies univariate analysis to the N-anomalies previously identified, ranks the results, and provides them to an AI to generate a diagnostics incident report. A prognostics module reviews the diagnostics incident report and maps each identified issue to a resolution plan. If execution of the resolution plan does not succeed in resolving the identified issue, the issue is escalated to a support team.

Abstract: Embodiments of the present technology provide memory cards intelligently designed to provide protection when mistakenly inserted into non-memory card hosts. Embodiments achieve such protection with less footprint/circuitry than existing fail-safe solutions, and without electrical overstress-causing offset voltages characteristic of existing fail-safe solutions. To realize these advantages, a memory card of the present technology includes a fail-safe reference voltage supply circuit that operates in a “fail-safe mode” by default, and exits and re-enters the fail-safe mode in response to voltage mode driver output enable (OE) signals.

Abstract: A system comprising processing circuitry a memory storing instructions that cause the system to detect a code change to source code included in a code repository, identify a relationship between the code change and an associated product feature, determine one or more dependent product features impacted by the code change, select a set of test cases including a subset of test cases related to the associated product feature and a subset of test cases related to the one or more dependent product features, execute the set of test cases, and update the code-to-feature mapping using results of executing the set of test case.

Abstract: A system and method for localizing an autonomous vehicle using mapped and real-time data. Mapped data and real-time data are scan matched. Characteristics of the autonomous vehicle such as, but not limited to, linear and angular velocities, heading, and motion prediction are provided to a Bayesian estimation algorithm, and a final pose is computed.

Abstract: Generally discussed herein are devices, systems, and methods for certificate management resiliency validation. A method can include receiving, at a resiliency tester of a cloud network, experiment data defining a certificate management resilience experiment to be performed on a resource of the cloud network, accessing, by the resiliency tester and a fault library and based on the experiment data, one or more faults that implement the certificate management resilience experiment, altering, by the resiliency tester and based on the experiment data, an attribute of a certificate, a version of the certificate, accessibility to a secrets vault that houses the certificate, or a combination thereof resulting in a modified certificate or modified certificate availability, and altering a certificate management tool based on telemetry data regarding access of the modified certificate.

Abstract: A charger for an electric vehicle includes a battery system; a high-voltage direct current (HVDC) bus configured to connect the charger to an external battery system; a DC-to-DC converter; and a switching component configured to (i) in a first operational state, connect the battery system to the HVDC bus directly so as to provide a parallel electric connection to the external battery system, and (ii) in a second operational state, connect the battery system to the HVDC bus via the DC-to-DC converter to align a voltage across the battery system with a voltage across the HVDC bus