Abstract: Described is a technology that facilitates directory level snapshots at any level of a filesystem's namespace. A directory inode mapping file associates a first directory's virtual inode number(s) with real inode number(s). A snapshot of the first directory creates a second directory with a second directory inode mapping file copied from the first directory inode mapping file, thereby sharing the real inode data structure(s) of the first directory. In the event that one of the directories is to be modified, the directory real inode is split to provide a new directory real inode file for the to-be-modified directory, with the modification after the split.

Abstract: Hairstyle transfer is challenging due to hair structure differences in the source and target hair. Latent Optimization of Hairstyles via Orthogonalization (LOHO) is an optimization-based approach using GAN inversion to infill missing hair structure details in latent space during hairstyle transfer. Hair is decomposed into three attributes: perceptual structure, appearance, and finer style, and includes tailored losses to model each of these attributes independently. Two-stage optimization and gradient orthogonalization enable disentangled latent space optimization of the three hair attributes. Using LOHO for latent space manipulation, users can synthesize novel photorealistic images by manipulating hair attributes either individually or jointly, transferring the desired attributes from reference hairstyles.

Abstract: Techniques are disclosed relating to merge operations for multi-level data structures, such as log-structured merge-trees (LSM trees). A computer system may store, in a database, a plurality of files as part of an LSM tree and a plurality of database key structures. A given one of the plurality of database key structures may indicate, for a corresponding one of the plurality of files, a set of key ranges derived from database records that are included in the corresponding file. The computer system may determine, using ones of the plurality of database key structures, a key range overlap that is indicative of an extent of overlap of key ranges from a set of the plurality of files with respect to a particular key range. Based on the determined key range overlap, the computer system may assign a priority level to a merge operation that involves the set of files.

Abstract: An exemplary method of imaging tissue of a subject using a rolling shutter imager to provide a video stream comprises: sequentially resetting a plurality of rows of pixels of the rolling shutter imager from a first row to a last row; transitioning a liquid crystal shutter from a closed state to an open state; after the liquid crystal shutter is transitioned into the open state and after resetting the last row, illuminating the tissue of the subject with an illumination light for an illumination period to accumulate charge at the plurality of rows of pixels, and after the illumination period ends, sequentially reading the accumulated charge at the rows of pixels from the first row to the last row; generating an image frame from the sequentially read accumulated charge at the plurality of rows of pixels; and adding the image frame to the video stream.

Abstract: An example system including a vehicle having a multi-zone network comprising a plurality of network zones, each one of the network zones comprising at least one zone manager; and a service management controller, including a service registration circuit structured to maintain a service register; a service publication circuit structured to expose at least one service; a service subscription circuit structured to enroll a requestor source; a network management controller, including a zone implementation circuit structured to interpret a zoned architecture communication scheme; a zone manager command circuit structured to interpret a zone manager command description; and a zone execution circuit structured to provide the corresponding zone manager command description to each of the at least one zone managers; and wherein each of the at least one zone managers is responsive to the corresponding zone manager command description and the service register to control communications.

Abstract: A method triggers a user-definable emergency response in a voice-command based authentication system. The method receives gestures on a mobile device, and provides an interface at the mobile device to trigger an automatic interaction with an artificial intelligence (“AI”) engine stored on the mobile device. The AI accesses a database for storing legacy information associated with the caller and runs an application program interface that provides the AI engine access to the legacy information. The AI engine receives the gestures generated by the caller, accesses the database storing legacy information generated by the caller and constructs an emergency profile of the caller. The emergency profile may be based on the legacy information, a machine learning model; and/or a user training session that initializes an emergency response signal threshold level. When the gestures exceed the emergency response signal threshold level, the method may generate the emergency response.

Abstract: An example method includes detecting, using sensors, packets throughout a datacenter. The sensors can then send packet logs to various collectors which can then identify and summarize data flows in the datacenter. The collectors can then send flow logs to an analytics module which can identify the status of the datacenter and detect an attack.

Abstract: Machine-learning-aided automatic taxonomy for web data. In an embodiment, a training dataset of annotated features is used to train a model to predict a class in a taxonomy of web-based activities. The features may be derived from a uniform resource locator (URL) of an online resource and associated metadata. During operation, the features may be extracted from the URL and metadata of each activity record in web data. The trained model may be applied to the extracted features for each activity record to predict a class within the taxonomy. The predicted taxonomic class may be stored in association with the URL that was extracted from the activity record to produce a taxonomized URL.

Abstract: A method for integrating a machine learning (ML) model that impacts different factor groups for generating a dynamic recommendation to collectively optimize an objective factor is provided.

Abstract: The disclosed technology provides solutions for providing an accurate pick-up location for an autonomous vehicle and a rider during a ride hailing pick-up. A method of the disclosed technology can include steps for determining map information associated with at least one thoroughfare for routing an autonomous vehicle, wherein the map information is based on location data associated with a client device requesting the autonomous vehicle; receiving sensor data from the client device corresponding to the at least one thoroughfare; determining, based on the sensor data and the map information, a first position of the client device relative to the at least one thoroughfare; and routing the autonomous vehicle based on the first position of the client device relative to the at least one thoroughfare. Systems and machine-readable media are also provided.

Abstract: Techniques for dynamically determining a configuration file are described. In an example, a system receives, from a device, a request for first content to be presented by the device. The system determines metadata of the first content and a cue point indicating a start of a segment in which second content can be presented by the device. Based on the metadata and subsequent to the request, the system determines a configuration of the segment. The configuration includes: a duration of the segment, a sequence of slots in the segment, and/or a content duration usable in a slot of the segment. The system generates a configuration file that includes the metadata and the configuration, determines the second content based at least in part on the configuration file, and sends data about the first content and the second content to the device.

Abstract: An adaptive prefetcher for a shared system cache of a processing system including multiple requestors having a cache miss monitor and a prefetch controller. The cache miss monitor monitors requests for information from memory and identifies one of the requestors for which an identified cache line is requested. The prefetch controller submits an adaptive request for a subsequent cache line. The subsequent cache line may be determined based on a latency comparison between a loop latency (LL) of the prefetch controller and a stream latency (SL) of the identified requestor. A latency memory may be included that stores stream latencies for the requestors. The latency comparison may be used to determine how many cache lines to skip relative to the identified cache line, such as according to SL*SK<LL?SL*(SK+1) in which SK is the number of cache lines to skip.

Abstract: Systems, methods, and computer program products are provided for efficiently storing multi-threaded log data. A method includes receiving multi-threaded log data comprising logs, markers, and thread identifiers. For each respective log, the respective thread identifier is set as a most recently used item in a thread reference cache. A respective log cache in a map data structure is determined based on the respective thread identifier. The respective log is added to the respective log cache. Whether to communicate the respective log and/or the respective log cache to a first repository is determined based on the respective marker. The respective log is communicated to a second repository. Whether to remove an oldest log from the log cache is determined based on a log cache size limit and/or a time limit. Whether to remove a least recently used log cache is determined based on at least one map data structure size limit.

Abstract: A battery charger includes a battery power regulator configured to set a charge signal provided to a battery based on a charge control signal and charge enable signal. The battery charger also includes a controller configured to provide the charge control signal to the battery power regulator. The controller is also configured to temporarily de-assert the charge enable signal for a predetermined amount of time in response to determining that a change is needed in the charge control signal. The controller is further configured to re-assert the charge enable signal after the predetermined amount of time.

Abstract: Battery management techniques for a vehicle include a set of sensors configured to measure a set of parameters of a battery of the vehicle and a controller configured to control recharging of the battery to a first target state of charge (SOC) corresponding to optimized battery life when a mileage of the vehicle is less than a threshold mileage corresponding to an expected transport period of the vehicle, wherein controlling the recharging of the battery to the first target SOC prevents battery malfunctions and thereby reduces vehicle warranty costs for an original equipment manufacturer (OEM) of the vehicle, and control recharging of the battery to a second target SOC determined by a cost-based optimization technique when the mileage of the vehicle reaches the threshold mileage.

Abstract: A method for time sensitive network (TSN) operation includes: in the case of data traffic stoppage in a communication channel of TSN due to at least one of down time of a user plane (UP) connection and a reset of the UP connection, performing the following: generating, by a user plane function (UPF), an audit report about the at least one of the down time and the reset; sending, by the UPF, the audit report to a first element in control plane (CP), without waiting for communication from the CP; and determining, by one of the first element or a second element in the CP, survival time of the communication channel. The audit report is sent along with a report type indicating the down time of the UP connection due to connection fluctuation or a switch-over of the UP connection from a primary UPF to a back-up UPF.

Abstract: A deployment request is received for a plurality of clusters within a plurality cloud computing environments. The deployment request relates to one or more operational actions for computing devices associated with the plurality of clusters. The plurality of clusters is associated with a single application programming interface (API). A set of standards are received for the plurality of clusters from the API. The plurality of clusters is configured according to the set of standards to create a configured plurality of clusters. The configured plurality of clusters are deployed within the plurality of cloud computing environments.

Abstract: Systems and methods for capturing sentiments and delivering elevated proactive user experience are disclosed.

Abstract: An active road surface maintenance system and method developed for connected vehicles with the aid of a mobility digital twin (MDT) framework. A method performed in a cloud-based digital space includes receiving data regarding a physical object from a physical space connected to a vehicle. The method also includes processing the data using machine learning to model road surface conditions, in which respective penalty values are assigned to corresponding road surfaces, a respective penalty value being higher the lower a condition of the corresponding road surface. The method also includes deriving instructions based on the modeled road surface conditions and the respective penalty values to guide actuation of the vehicle along a trajectory. The method further includes transmitting the instructions to the physical space connected to the vehicle to guide actuation of the vehicle.

Abstract: Certain aspects of the present disclosure provide techniques for handover timelines and power optimization for user equipment (UE) in multiple subscriber identity module (MSIM) dual receive (DR) scenarios. An example method, performed by a UE, includes entering a DR mode in which a first subscriber identity module (SIM) is active in a first radio access technology (RAT) and a second SIM is idle in a second RAT, identifying one or more channel frequencies, of a neighbor channel frequency list, that are incompatible with the DR mode, and performing a channel measurement procedure with the second SIM, on a subset of the neighbor channel frequency list that does not include the one or more channel frequencies that are incompatible with the DR mode.