Microsoft Patente

The techniques describe dynamic content-aware data sampling that adjusts a sampling rate of data associated with a service based on the content of the data. More specifically, if the data to be logged indicates that the service is operating normally, then the sampling rate is adjusted lower as the logged data has limited purpose provided the service is likely not experiencing an issue. In contrast, if the data indicates that the service is operating abnormally, then the sampling rate is adjusted higher as the logged data has elevated purpose provided the service may be experiencing an issue. The dynamic content-aware data sampling is implemented using program code generated by an artificial intelligence agent (e.g., a generative language model) based on service owner requirements, as well as historical data patterns and/or usage patterns.

Information is received indicating a plurality of files that are determined to be associated with a user's computing activity in the computing system and meet a criterion for being variants of one another. A polymorphic placeholder is generated that corresponds to the plurality of files. The plurality of files are selected based on a threshold relevance to the criterion, and the plurality of files are abstracted from underlying storage locations in the computing system.

A data processing system with a processor and memory is configured to: implement a feature pipeline to receive telemetry data generated in response to activity in a set of applications; analyze the telemetry data using a feature extraction unit of the feature pipeline to extract feature data from the telemetry data; operate a scoring pipeline to classify the user as working onsite or remotely based on the feature data extracted from the telemetry data, the scoring pipeline utilizing a machine-learning model trained to receive feature data as an input and to output a classification indicating that the user is working onsite at a location associated with the enterprise or remotely at a location not associated with the enterprise; and allocate computing and network resources using a resource allocation unit based on the classification.

Methods, apparatuses, and products for database query acceleration. Methods include receiving, by a query execution platform, database query plan data from a database host, the database query plan data comprising a query execution plan and one or more database tables, translating the database query plan data into format associated with a coprocessor of the query execution platform, generating a coprocessor query execution plan for the coprocessor based on the query execution plan of the database query plan data, identifying one or more operators for executing the coprocessor query execution plan, selecting an implementation for each of the one or more operators based on one or more characteristics of data on which the one or more operators are to be applied based on the coprocessor execution plan, and executing the coprocessor query execution plan using the coprocessor.

A cooling system (100) for a set of solid state drive (SSD) modules (200) includes a thermal mounting frame (110). The thermal mounting frame includes a thermal transfer plate (112), and a plurality of parallel fins (114) that project outward from a first side (116) of the thermal transfer plate along a length dimension (118). The first side of the thermal transfer plate in combination with each neighboring pair of fins (122) of the plurality of parallel fins defines a slot (124) that accommodates a corresponding SSD module (210) of the set of SSD modules. The thermal transfer plate includes a plurality of parallel spines (320) in which each spine projects into a core region (132) of a corresponding fin of the plurality of parallel fins along a base portion (310) of the length dimension of the fin. For each slot, each fin of the neighboring pair of fins that defines the slot includes an exterior thermal transfer layer (130).

A method for recommending a mixed-precision data type configuration to accelerate execution of a language model includes identifying a plurality of mixed-precision test configurations that assign mixed-precision data types among components of the language model; measuring saliency metrics for the language model while the language model is executed in a test environment configured in accord with multiple of the plurality of mixed-precision test configurations; generating a saliency ranking that ranks the components based on the saliency metrics; and generating the recommended mixed-precision data type configuration. The recommended mixed-precision data type configuration assigns a first, reduced-precision data type to a least salient subset of the components identified in the saliency ranking while assigning a second, higher-precision data type to a most salient subset of the components identified in the saliency ranking.

Systems and methods for generating synthetic data. One such method includes receiving, at a security unit, adaptive controls comprising both privacy budget constraints and compliance requirements for determining risk scores for sensitive information. The method further includes receiving, at the security unit, an input dataset comprising a plurality of data entries including sensitive information and determining a risk score in accordance with the compliance requirements for each of the plurality of data entries. The method further includes dynamically allocating, for each of the plurality of data entries, a privacy budget based on the risk score for the data entry and the privacy budget constraints. The method further includes performing adaptive remediation on the input dataset to generate a synthetic dataset, wherein adaptive remediation comprises adaptive aggregation and/or adaptive differential noise injection.

A two-die integrated circuit (IC) package is coupled to dual in-line memory module (DIMM) connectors. The DIMM connectors are homogeneously-oriented (e.g., in a same direction) but the dies on the IC package are oriented opposite to each other. The DIMM connectors are homogeneously oriented to avoid issues with memory card insertion and to minimize DIMM connector density. Because the DIMM connectors are coupled to both of the dies on the IC package, and the dies are oriented opposite to each other, the sub-channels of the physical interfaces (PHYs) on a second die of the two dies is oriented opposite to the sub-channel interfaces in the DIMM connectors. The second die directs memory transactions of a first sub-channel through a second sub-channel interface driver in the PHY and the other die directs memory transactions of a first sub-channel of the memory through a first sub-channel interface driver in the PHY.

A kinetic inductance travelling-wave parametric amplifier (KTWPA) device is provided that includes a metallic enclosure (302) having a base including an inside top surface, sides, and an enclosure lid (308), a substrate (402) coupled to the inside top surface of the base, the substrate including a plurality of conductive pads (206), a silicon die (202) including a superconducting transmission line (200), the silicon die coupled to the substrate, a plurality of conductive spacers (210) coupled to the silicon die, wire bonds (208) attached between the conductive pads and the silicon die, and a copper lid (212) coupled to the top end of the plurality of conductive spacers, where the copper lid is electrically and thermally coupled to the enclosure lid. The KTWPA device may be applied in a quantum microwave measurement circuit to amplify a quantum measurement signal from a quantum circuit.

Various embodiments relate to a method implemented in a computer system with a processor system, involving loading a container image as a first read-only filesystem volume for a virtual machine's guest operating system (OS). This includes obtaining a root integrity hash, validating a first hash tree within the container image, and ensuring data integrity. A virtual disk image is loaded as a second read-write filesystem volume, utilizing an ephemeral encryption key stored in the VM's memory. A filesystem is created with encryption and integrity protection features, enabling data encryption and hash-based validation. A third filesystem volume is created as a union of the first and second volumes, directing writes to the second volume. This third volume is exposed as a system volume for the guest OS, ensuring secure and validated data handling within the virtual environment.

Methods, apparatuses, and products for secure peripheral device state transfer during trusted virtual machine migration, including: receiving, by a first instance of a trusted virtual machine and from a first peripheral device coupled to a first host executing the first instance of the trusted virtual machine, a first hash value calculated based on a state of the first peripheral device; loading, by the first instance of the trusted virtual machine, a second hash value calculated based on a state of a second peripheral device coupled to a second host; and trusting, by the first instance of the trusted virtual machine, the first peripheral device based on a match between the first hash value and the second hash value.

The present systems and methods enable a web browser to recommend or present features to a user based on the content of a web page. The features, which may include various actions that can be taken and different display configurations for viewing the content of the web page, may be specific to the type(s) of content that are present on the web page. Furthermore, such features may be presented to the user upon detection of a trigger that indicates an intent to arrange or configure a tab of the browser. Such features may be presented as transient graphical elements that illustrate the feature.

This document generally relates to USB-C connections between devices and cords or cables. One example device includes a housing including multiple surfaces that collectively define a volume. The device also includes a USB-C receptacle positioned within the volume and recessed behind an individual surface. The recessed USB-C receptacle includes a USB-C compliant arrangement of pins. The example also includes a funnel that extends through the individual surface to the recessed USB-C receptacle.

An improved kinetic inductance travelling-wave parametric amplifier (KTWPA) device is provided that includes a first substrate (206), a superconducting trace (204) having a first height (H1), the superconducting trace fabricated on the first substrate, a plurality of spacers (208) having a second height (H2), each of the plurality of spacers fabricated on the first substrate, the second height greater than the first height of the superconducting trace, a ground plane (210) bonded to a second substrate (212), the ground plane further bonded to the plurality of spacers, and a vacuum gap (202) formed between the first substrate and the ground plane. A superconducting transmission line assembly (200) is formed by the superconducting trace and the ground plane separated by the vacuum gap having the second height. The KTWPA device is applied in a quantum microwave measurement circuit for amplifying weak quantum microwave signals at a very low noise floor.

Circulators with voltage-tunable inductors formed using superconductor-semiconductor junctions are described. An example circulator includes a first transfer switch (220) including a first plurality of ports, where the first transfer switch comprises a first plurality of superconductor-semiconductor junctions. The circulator further comprises a second transfer switch (240) including a third plurality of ports, where the second transfer switch comprises a second plurality of superconductor-semiconductor junctions. The circulator further comprises a first delay line (212) and a second delay line (214) coupled between the first transfer switch and the second transfer switch. Modulation of gate voltages being applied to respective superconductor-semiconductor junctions allows the first transfer switch and the second transfer switch, in combination with the first delay line and the second delay line, to enable non-reciprocal communication with the quantum device.

An apparatus and method for mitigating gas line absorption of a signal transmitted via a hollow core fiber by a transmitter are described. The method comprises computing an adjustment to be made to a property of the signal, the adjustment being an adjustment to change one or more of a frequency, an amplitude, and a modulation format of the signal to reduce an impact of gas line absorption on the signal. The method further comprises configuring one or more of: the transmitter to cause the adjustment to be applied to the signal during transmission of the signal, via the hollow core fiber, by the transmitter; and a receiver configured to receive the signal transmitted via the hollow core fiber, to cause the adjustment to be applied to the signal after the receiver receives the signal.

An apparatus and method for mitigating gas line absorption of a signal transmitted via a hollow core fiber by a transmitter are described. The method comprises initiating generating a signal using a laser and initiating modulating the signal to produce a carved signal, wherein the carved signal comprises a plurality of channels. A spectrum of the carved signal comprises substantially zero signal intensity at a frequency within a gas line absorption profile, the frequency being between a first channel and a second channel of the carved signal. The method further comprises initiating transmission of the carved signal via the hollow core fiber by the transmitter.

A computing system generates a trained super-resolution model using correlated high-resolution and low-resolution image pairs. A first image generator produces high-resolution images, which are processed by a downscaler network to create corresponding low-resolution images. The downscaler network is trained using an initial set of image pairs. Additional high-resolution images are generated and processed by the downscaler network to produce additional low-resolution images, forming a second set of image pairs. The super-resolution model is trained using the second set of image pairs. The trained super-resolution model processes low-resolution images to generate enhanced images with improved detail and reduced artifacts. The model can be used to increase resolution of input images or refine details without changing resolution. The trained model may be deployed on different devices and applied across various rendering environments.

In various examples there is a computer-implemented method, comprising ingesting into a memory a batch of media items. Duplicate media items in the batch are identified. For each duplicate media item, the duplicate media item is ranked with respect to duplicates of the media item. Filtering the duplicate media items is performed using the ranks of the duplicate media items to obtain filtered media items. Metadata of individual ones of the filtered media items is computed. Individual ones of the filtered media items are labelled with part of the metadata. For each filtered media item having metadata and a label, input into a Large Language Model, LLM, the media item, the label, the metadata and a request, and receive from the LLM a description of content depicted in the media item. A graph database of the filtered media items is created using the descriptions, the metadata, and the labels.

A computerized method processes network packets in parallel. Packets are obtained from network traffic of a network tunnel. The packets are distributed between two input queues, wherein each input queue is associated with a dedicated packet processing branch that can operate in parallel with other packet processing branches. Packets from the input queues are processed using dedicated packet processors (e.g., dedicated CPU core(s) and/or other dedicated processing entities) and the resulting processed packets are added to associated output queues. Packet order information is used to maintain the proper packet order across the multiple packet processing branches such that, when the processed packets from the multiple output queues are combined into a single processed packet flow or stream, the packets are in the same order as when they entered the network tunnel. After the processed packets are combined, they proceed along the network tunnel.

Systems, methods, and computer program products are disclosed for a multi-level workload flighting service. A received customer job request for a customer job is copied to create a shadow job request for a shadow job. The customer job request is provided to a first service having a first configuration, and the shadow job request is provided to a flighting service. The flighting service determines that the shadow job request satisfies a flighting criteria and provides the shadow job to a second service having a second configuration. Execution data resulting from execution of the customer job and the shadow job are collected for analysis.

The present disclosure generally relates to systems and methods for automatically inferring implicit relationships between columns of an input table. For example, the systems and methods discussed herein infers semantically-accurate functional inter-column relationships with a unified framework that includes speed up options. The systems and methods discussed herein leverage principled statistical tests based on hypothetical analysis that slightly "perturbs" values in an input table to determine the strength of a potential functional relationship between some subset of the input table columns. Based on this determined strength, the systems and methods discussed herein reject potential functional relationships that are likely false-positive while retaining potential functional relationships that are likely true-positive.

Systems and methods for dynamic customization for object detection. An object detector uses an object detection model (e.g., an open vocabulary model) to detect and classify objects. The object detector supports customization of classifications without requiring the model to be retrained. Embeddings are used to differentiate between various states of detections, enhancing accuracy and relevance of the detection process. In examples, the object detector handles logical conditions by creating a classifier trained to differentiate between various combinations of objects. Aspects of the present disclosure support the use of images for customization, allowing users to provide examples from which the object detector described herein can learn. In further examples, capabilities of the object detector extend to processing live video streams, where users can mark examples in real-time. In some examples, the object detector leverages a generative model (e.g., to provide state contrasts), allowing the system to be adaptable to complex conditions.

This disclosure describes utilizing an anomaly-based mitigation system to accurately, flexibly, and efficiently determine groupings of anomalous signals from across a computing system by identifying and combining different logical groups of anomalous signals. For example, the anomaly-based mitigation system utilizes different grouping logic models, which identify various connection strategies to determine initial groupings of anomalous signals. The anomaly-based mitigation system then determines commonalities between the initial anomalous signal groups using graph models to identify connected groups of anomalous signals. Furthermore, the anomaly-based mitigation system refines the groups to enhance their effectiveness in identifying and resolving service incidents in the computing system. In some instances, the anomaly-based mitigation system presents the anomalous signal groups in an incident report interface that includes actions to take or actions that were automatically taken to resolve a service incident.

The present disclosure relates to systems and methods for identifying cybersecurity risks in a cloud environment. The systems and methods generate a unique fingerprint for an event in the cloud environment using a hash structure. The systems and methods detect an anomaly in the event using the fingerprint. The systems and methods use the anomaly in identifying a cybersecurity risk in the cloud environment.

Packets are captured that are associated with a communications flow in a communications network. An image generated by converting bits of the captured packets to pixel data. The image is input to a machine learning model that is trained to identify an operating system or operating system version based on image data. Identities of operating systems or operating system versions of source devices of the packets are received. Based on the identities, it is determined whether the packets are associated with at least two different operating systems or operating system versions. This indicates the use of device tethering in the communications flow.

An input device comprises a touch surface and a printed circuit board comprising a switch. A mounting bracket comprises a boss aligned with the switch, and a plurality of flat springs each comprise a bracket end affixed to the mounting bracket and an opposite end coupled to the printed circuit board.

A method disclosed herein provides for discovering and extracting database schema information from a database; generating a database graph that graphically depicts the database schema information; generating contextual data that includes or that is derived from the database graph; presenting a language model with the contextual data and an instruction to use the contextual data to generate schema element descriptors for the schema elements; storing the schema element descriptors generated by the language model based on processing of the instruction; and making the stored schema element descriptors available to a query translation system that performs query translation tasks.

Methods, apparatuses, and products for leveraging a large language model ('LLM') for modernizing a codebase, including: generating, using a codebase as input, a hierarchical representation of the codebase, wherein the hierarchical representation of the codebase includes one or more summarizations of functional units within the codebase and information describing relationships between the functional units within the codebase; and generating a searchable index that includes vector representations for nodes in the hierarchical representation of the codebase, wherein the LLM generates a response to a query based on content contained in the searchable index.

Aspects of the disclosure include a method for determining and correcting an error in measurement outcomes from a quantum circuit. Aspects include receiving, from the quantum circuit, output of the measurement outcomes for a four-dimensional (4D) toric code and determining a flux loop from the measurement outcomes, where the flux loop corresponds to a 4D torus. Aspects include transforming the flux loop into a 4D covering space by fixing a base point of the flux loop in the 4D covering space, where a 4D bounding box is generated that encloses the flux loop in 4D covering space. Aspects include shrinking the 4D bounding box enclosing the flux loop by decreasing a size of the bounding box, where the shrinking continues until the 4D bounding box surrounds a single plaquette, where a pattern of errors correspond to the single plaquette. The pattern of errors corresponding to the single plaquette is corrected.

Automated deployment and maintenance of a cloud fleet, including: receiving a request to generate a cloud fleet in a cloud computing environment, the request including a listing of selectable cloud computing instance configurations, an allocation strategy, and a resource allocation; selecting, for each cloud computing instance configuration of one or more cloud computing instance configurations included in the listing of selectable cloud computing instance configurations, based a capacity of the cloud computing environment, a corresponding number of instances to satisfy the resource allocation, including selecting one or more cloud computing instance configurations from the listing of selectable cloud computing instance configurations based on the allocation strategy; and creating based on the corresponding number of instances for each cloud computing instance configuration, the cloud fleet comprising at least one of: one or more on-demand cloud computing instances and one or more spot cloud computing instances.

A quantum computing system (1) including a quantum computing device (10) that includes a plurality of qubits (14). The quantum computing system further includes a qubit interface (34) including a readout circuit (100). The readout circuit is configured to collect a qubit measurement readout (40) of one or more of the plurality of qubits. The qubit measurement readout further includes a signal-frequency readout (42) collected at a signal frequency (fs) and an idler-frequency readout (44) collected at an idler frequency (fi). The quantum computing system further includes a classical computing device (18) configured to receive the qubit measurement readout from the readout circuit. The classical computing device is further configured to compute a weighted sum (50) based at least in part on the signal-frequency readout and the idler-frequency readout. The classical computing device is further configured to output a combined measurement readout of an observable (52) of the one or more qubits, as indicated by the weighted sum.

Systems and methods are disclosed herein for improvements to memory registration. In an example system, a set of page table entries (PTEs) are reserved in a first network interface device of a first computing device. In the example, each page table entry comprises a placeholder for a memory pointer. A page in the physical memory is pinned by a kernel of the first computing device. In one example, the request to pin the page in the physical memory originates from an application seeking to perform an RDMA operation. A page table entry of the set of PTEs in the network interface device is configured with information indicative of a physical memory address corresponding to the pinned page. A remote direct memory access operation is subsequently performed over a network based on the page table entry.

Methods, apparatuses, and products for automatic deployment of a multi-region cloud fleet, including: receiving a request to generate a cloud fleet in a cloud computing environment, wherein the request comprises: data indicating a plurality of selectable regions for the cloud fleet and a resource allocation indicating an amount of computational resources to be allocated for the cloud fleet; determining, based at least on a capacity of the plurality of selectable regions, a cloud computing instance distribution across at least one region of the plurality of selectable regions and satisfying the resource allocation, wherein the cloud computing instance distribution describes, for each region of the at least one region, a corresponding group of cloud computing instances to be created for the cloud fleet; and creating, in each region of the at least one region, the corresponding group of cloud computing instances described in the cloud computing instance distribution.

Regenerable items serve as identifiers or keys or both. A regenerable items tuple is generated by combining context-specific identifiers into a string, hashing the string, and designating individual hash portions as respective regenerable items. Regenerable items maintain artifact tracking continuity after a reinstall, a reboot, or another intervening event without retrieval of a previously generated value from a registry or state file. Some regenerable items conform to a universally unique identifier specification. Some regenerable items are free of reliance on nondeterministic content, and some are free of reliance on time-dependent content.

Techniques are described herein that are capable of causing an AI system to convert a legacy application into a converted application using a schema. An AI system is caused to generate a schema that defines functional blocks of a legacy application and logical relationships therebetween based on documentation describing the functional blocks and the logical relationships. The AI system is caused to perform a conversion of the legacy application into a converted application based on the schema and the legacy application. Performing the conversion includes converting the functional blocks, which are configured to have respective functionalities and are written in a first programming language, into code blocks, which are configured to have the respective functionalities and are written in a second programming language. Performing the conversion further includes configuring the code blocks to have the logical relationships. The converted application is defined by the code blocks and the logical relationships.

Methods, computer systems, and computer storage media are provided for facilitating implementation of efficient management of monitor configuration data in a remote desktop environment. In embodiments, a dynamic virtual channel established between a remote desktop client device and a remote desktop server is used to provide a global set of monitor configuration data for a remote desktop session. Thereafter, at the remote desktop client device, an updated set of monitor configuration data is identified during the remote desktop session. The updated set of monitor configuration data for the remote desktop session is provided to the remote desktop server via the dynamic virtual channel.

Methods, apparatuses, and products for cardinality estimation feedback for expressions in a database query, including: calculating, for one or more expressions of a database query, a default estimated cardinality and an actual cardinality, wherein the default estimated cardinality is based on an assigned cardinality estimation (CE) model comprising a default CE model, wherein the actual cardinality is based on an execution of the database query; calculating, for the one or more expressions of the database query and based on the default estimated cardinality and the actual cardinality, a default cardinality error; selecting, for a particular expression of the one or more expressions of the database query, based on the default cardinality error for the particular expression exceeding a first threshold, a recommended CE model; and assigning the recommended CE model as the assigned CE model for the particular expression based on an improvement in one or more respective recommended cardinality errors for the particular expression relative one or more respective default cardinality errors for the particular expression exceeding a second threshold, wherein the one or more respective recommended cardinality errors for the particular expression are based on one or more recommended estimated cardinalities calculated using the recommended CE model.

Systems and methods are provided for a novel algorithm and system for automatically estimating the loss in an optical circuit switches ("OCS") network and identifying components with unusually high loss. A computing system receives a first set of optical power data corresponding to optical power emitted from each of a plurality of optical transmitters that is communicatively coupled to input ports of an OCS, via optical fiber cables, while receiving a second set of optical power data measured by a photodetector of each of a plurality of optical receivers that is communicatively coupled to output ports of the OCS, via optical fiber cables. The computing system calculates a plurality of optical signal losses, based on analysis of the first and second sets of optical power data in view of expected ranges of optical power. The computing system generates a report, based on the plurality of optical signal losses.

Methods, apparatuses, and products for trust establishment for cloud computing environments include establishing an initial trust between an intermediary component of a computing environment and a host of the computing environment; intercepting, by the intermediary, a request from software executing within the computing environment to an identity authentication endpoint; determining, by the intermediary, whether to authenticate the software; in response to determining to authenticate the software, applying, by the intermediary, a digital signature to the request; and forwarding the request, with the digital signature, to the identity authentication endpoint.