Microsoft Patente

Technologies are disclosed for providing visual feedback during touch-based operations on user interface ("UI") elements. Through implementations of the disclosed technologies, visual feedback can be provided to users during certain touch-based operations on UI elements presented by touch-enabled computing devices. The visual feedback can provide confirmation to users that certain touch-based operations on UI elements were performed successfully. The visual feedback provided by the disclosed technologies can make it easier for users to select a UI element, to understand that actions taken to select a UI element were successful, and to successfully perform operations on UI elements once they have been selected. These benefits can reduce the number of times users need to attempt certain touch-based operations on UI elements. This, in turn, can reduce the utilization of computing resources, such as memory and processor cycles, by touch-enabled computing devices implementing the disclosed technologies.

The present disclosure generally relates to offloading the orchestration and control of a communication collective to a specialized collective offload engine. The systems, methods, and specialized computing hardware described herein avoid the latencies and other inefficiencies introduced when software applications control the execution of a communication collective. For example, the described systems, methods, and specialized computing hardware generate a binary representation of one or more direct acyclic graphs of node operations for a communication collective, and load this representation into the instruction memories of specialized collective offload engines residing on network endpoints connected to one or more network switches. During execution of the collective, the collective offload engines initiate node operations based on whether corresponding dependencies are met. By offloading these tasks to the collective offload engines, no additional latencies are introduced and other computing resources are kept free for use by other applications.

Systems, methods, and techniques are directed to data operations utilizing ring data-structures. A device comprises a network interface controller (NIC) and a ring data-structure comprising a first ring comprising pointers indicating memory regions that are empty and a second ring comprising pointers indicating regions that store data. In an aspect, the NIC receives a write request and determines a region to write data to based on the first ring. The NIC writes the data to the region and updates the second region to include a pointer to the region. In another aspect, the NIC receives a read request to read data and determines a region to read data from based on a pointer in the second ring. The NIC reads the data from the region. In a further aspect, the NIC is a one-sided remote direct memory access (RDMA) NIC and reads/writes utilizing low-level operations.

The disclosed techniques provide dynamically controlled notifications that improve inclusiveness in online meetings. The dynamically controlled notifications can be audio or visual indicators that supplement a first notification. A user input causes the display of the first notification, an image of a "virtual hand raise." The input also invokes a timer for measuring an elapsed time from the time of the user input. The system also analyzes audio streams to detect topic transitions of a discussion of the meeting participants. Then, the system generates dynamically controlled notifications in response to at least one triggering event, either (1) in response to detecting that the timer has reached a time threshold or (2) in response to a detection of that the discussion has transitioned to a new topic. In some embodiments, the dynamically controlled notifications are invoked in response to either of the two triggering events, whichever event is detected first.

A system having scalable sockets to support User Datagram Protocol (UDP) connections identifies a plurality of UDP connections, wherein a plurality of remote clients connect to corresponding ones of the plurality of UDP connections. Each one of a plurality of UDP sockets is associated with a corresponding one of the plurality of UDP connections. A network stack lookup for UDP packets in network traffic is performed using a network stack to identify the UDP socket corresponding to the remote client associated with each of the UDP packet. The UDP packets are buffered with a send buffer and a receive buffer for the UDP socket corresponding to the remote client associated with the UDP packets as determined by the network stack lookup to support communication over the plurality of UDP connections using the plurality of UDP sockets. The system thereby operates more efficiently and/or is more scalable.

A method of dicing a semiconductor wafer comprises steps of a) positioning a stencil mask over the semiconductor wafer, with the stencil mask being spaced from the semiconductor wafer by a gap such that the stencil mask does not touch a component arranged on the semiconductor wafer, the stencil mask having an aperture; and b) subsequently etching through the semiconductor wafer by performing a reactive ion etch, wherein the stencil mask controls the etch. The method may allow a wafer to be diced without the use of die attach tape and without need for a cleaning step, and may be useful for dicing wafers bearing microelectromechanical systems. Also provided are a stencil mask and system useful in the method.

The technology described herein enables a remote desktop application to interact simultaneously with peripherals associated with multiple client devices. The technology described herein expands peripheral options to include peripherals associated with a user's other client devices that have a desktop application installed. For example, a user may access a virtual machine on a primary client, but use the touchscreen on a secondary client to control the virtual machine. The primary and secondary clients may have different connection types with the remote virtual machine. The connection types include a full connection and a background connection. In addition to different connection types, the remote desktop applications on the primary and secondary clients may operate in different modes. Two different modes include graphics mode and non-graphics mode.

A hardware device is communicatively connected to host servers individually supporting a virtual switch and one or more virtual machines and includes a processor input to receive a header of a data packet in the hardware device, a hardware rules processor scheduler to schedule processing of the header among hardware processing cores in the hardware rules processor, a rules processor core scheduled by the hardware rules processor scheduler to receive the header from the hardware rules processor scheduler, the rules processor core to read compiled networking policies from the memory, and to apply the compiled networking policies to the header to yield a modified header for the data packet, and a connection establishment processor to record the networking connection for the data packet based on the modified header in an established connection datastore.

Systems, methods, devices, and computer readable storage media described herein provide techniques for generating data queries. In an aspect, a request associated with querying a service is received. The request comprises a natural language (NL) question. A prompt comprising the NL question is provided to a generative artificial intelligence (AI) model to cause the model to determine additional information is to be determined to translate a language of the NL question to a query language, utilize a function call of the service to determine the information, and generate a data query based on the information. The data query is received from the model and provided to the service. In an embodiment, the prompt causes the model to generate a query plan.

Systems, methods, and computer program products are disclosed for scalable event stream processing of trigger rules. Event data associated with a set of object is partitioned into a set of logical partitions, including a first logical partition comprising first event data associated with a first object. The first logical partition is provided to a first event processor to enable the first event processor to update a state associated with the first object based on first event data in the first logical partition. Satisfaction of a first trigger condition specified by a first trigger rule is determined based on the state associated with the first object. A first action is performed in response to the satisfaction of the first trigger condition.

A method for battery authentication includes, at a battery controller of a battery subsystem, measuring a test electrochemical impedance spectroscopy (EIS) response of one or more battery cells connected to the battery controller. The test EIS response is compared to a historical EIS response stored in a storage subsystem, wherein the historical EIS response corresponds to one or more authentic battery cells associated with the battery controller. Based at least in part on a measured similarity between the test EIS response and the historical EIS response, the battery controller outputs an authenticity determination for the one or more battery cells connected to the battery controller.

An integrated circuit and method are disclosed for processing of compressed columnar data. The processing of compressed columnar data includes loading compressed columnar data associated with a first and a second column into a first on-chip buffer, performing time-shared processing of the compressed columnar data from the first on-chip buffer into a second on-chip buffer, transcoding the compressed columnar data from the second on-chip buffer into unified columnar data having a unified format, loading the unified columnar data into a third on-chip buffer so that the unified columnar data are logically aligned in the third on-chip buffer, and providing at least a portion of the unified columnar data to a query operator. The integrated circuit includes a column loader, a balancer, a transcoder, and a decoder for performing processing of the compressed columnar data.

The techniques disclosed herein provide a system for segmenting a sequence of content captures (e.g., screenshots) of a desktop environment based on a semantic relationship between individual content captures. Generally described, the system generates a numerical representation (e.g., an embedding) of a content capture in the sequence. The numerical representation is then compared against numerical representations of neighboring content captures to detect changes in user activity such as switching activities. Accordingly, the system calculates a difference metric that quantifies the level of change between content captures and compares these difference metrics against a threshold difference metric to identify such changes in user activity. In the event at least one difference metric satisfies the threshold difference metric, the system partitions the sequence of content captures to generate at least a first segment and a second segment. The segments are then rendered in an interactive timeline interface.

A computerized system and method for flow orchestration for language model-based agents are provided. A workflow comprising a plurality of agents configured to execute in a multistep multi-pass (MSMP) mode is defined. A request for data is received by a generative artificial intelligence (GAI) model. A portion of the requested data is retrieved based on executing a first agent of the plurality of agents of the workflow e.g., in a first pass of the MSMP mode. The GAI model adjusts the workflow based on the retrieved portion of the requested data. For example, an order of the plurality of agents of the workflow (e.g., which agent is to be executed first, second, and so on) is adjusted based on the retrieved portion of the requested data. The requested data is obtained based on executing the adjusted workflow.

A processing system comprising one or more chips, each comprising a plurality of tiles is described. Each tile comprises a respective processing unit and memory, the memory storing a codelet. The processing system has at least one encryption unit configured to encrypt and decrypt data transferred between the tiles and a trusted computing entity via an external computing device. The codelets are configured to instruct the tiles to transfer the encrypted data by reading from and writing to a plurality of memory regions at the external memory such that a plurality of streams of encrypted data are formed, each stream using an individual one of the memory regions at the external computing device.

Systems and methods herein provide an interface engine and its related functions. In an example, a method includes identifying, by an interface engine, an event involving a peripheral for a client device. The peripheral is an external device that extends the functionality of the client device. Responsive to identifying the event, the interface engine determines connection capabilities of the client device and determines connection for the peripheral based on the connection capabilities of the client device and the event. The interface engine may then generate and send a notification of the connection to the client device.

Implementations for simplified and efficient error correction code schemes with multiple codewords are provided. One aspect includes a computing system comprising: processing circuitry and memory storing instructions that, during execution, causes the processing circuitry to: encode data comprising user data and metadata in a memory module comprising a plurality of dies by: logically partitioning the memory module into a plurality of partitions; on a first partition: storing a first portion of the user data on one or more memory segments of the first partition; and storing a first cyclic redundancy check (CRC) codeword and the metadata on a memory segment of the first partition different from the one or more memory segments of the first partition, wherein the first CRC codeword is generated based on the first portion of the user data and the metadata; and storing parity data on one of the plurality of dies.

A method for secure key exchange. The method comprises receiving a request to certify a key from a communication partner at an interface between an access and tamper resistant circuit block and exposed circuitry. Within the access and tamper resistant circuit block, a first random private key is generated. A corresponding public key of the first random private key is derived, and a cryptographic digest of the public key and attributes associated with the first random private key is generated. The generated cryptographic digest is signed using a second random private key that has been designated for signing by one or more associated attributes. The public key and the signature are then sent to the communication partner via the interface.

A method for exciting a resonator at an estimated resonance frequency of the resonator comprises: (a) varying a frequency of an input signal applied to the resonator while recording an output signal from the resonator; (b) for each of a series of frequency values of the input signal, computing noise-filtered, complex first and second derivatives of a ratio of the output signal to the input signal; (c) for each of the series of frequency values of the input signal, computing a real part of a dot product of the complex first and second derivatives as vectors projected onto a complex plane; and (d) exciting the resonator an excitation frequency where the real part vanishes.

A system may include a memory device to store instructions and data, and at least one processing device to execute the instructions stored in the memory device to: receive a background image and a digital object to be composited onto the background image in a mixed reality view, generate a 2D bounding region for the digital object, select a version of the background image at a pre-defined resolution, overlay the 2D bounding region on the selected version of the background image and obtain a set of samples of the colors of pixels of the selected version along a perimeter of the 2D bounding region, and determine a value for one or more digital lighting sources to illuminate the digital object in the mixed reality view, based, at least in part, on the set of samples.

A system may include an electrical switch. A system may include a heat exchanger in thermal communication with the electrical switch and configured to receive heat from the electrical switch. A system may include a liquid cooling conduit configured to flow a working fluid therethrough, wherein the liquid cooling conduit is in thermal communication with the heat exchanger and configured to receive heat from the heat exchanger. A system may include an exhaust device in fluid communication with the liquid cooling conduit and configured to exhaust heat from the working fluid.

This application relates to video encoding and decoding, and specifically to tools and techniques for using and providing supplemental enhancement information in bitstreams. Among other things, the detailed description presents innovations for bitstreams having supplemental enhancement information (SEI). In particular embodiments, the SEI message includes picture source data (e.g., data indicating whether the associated picture is a progressive scan picture or an interlaced scan picture and/or data indicating whether the associated picture is a duplicate picture). The SEI message can also express a confidence level of the encoder's relative confidence in the accuracy of this picture source data. A decoder can use the confidence level indication to determine whether the decoder should separately identify the picture as progressive or interlaced and/or a duplicate picture or honor the picture source scanning information in the SEI as it is.

Methods, apparatuses, and products for accelerating queries using a multiple graphics processing units ('GPUs') in a GPU cluster, including: generating, by each GPU in the GPU cluster, a local representation of a portion of a dataset that is stored in memory of the GPU, wherein the entire dataset is distributed across the GPU cluster; broadcasting, by each GPU in the GPU cluster, its local representation to one or more other GPUs in the GPU cluster; and generating, by each GPU in the cluster and based on local representations received from the other GPUs in the GPU cluster, a global representation of the dataset; and transferring data elements stored in memory of a first GPU to one or more other GPUs in the GPU cluster, wherein the data elements are selected using the global representation of the dataset.

Methods, apparatuses, and products for securely deploying cloud-native applications, including: creating, within a tenant's cloud deployment, a secure execution environment for an application, wherein the secure execution environment is managed exclusively by a cloud service provider; deploying, within the secure execution environment, the application, wherein source code for the application is stored in the secure execution environment; and deploying an agent within the secure execution environment, wherein the agent is configured to allow one or more conforming requests to access the application, block one or more tenant-initiated management operations for the secure environment, and allow one or more management operations for the secure environment that are initiated by the cloud service provider.

Systems and methods are provided for detecting memory hazards in memory, such as in massively parallel and distributed systems. In examples, a memory hazard logic detects access of a memory by a device communicatively coupled to the memory. The memory hazard logic identifies a memory access type and a memory address range being accessed, based on information obtained when monitoring a memory access transaction between the device and the memory. The memory hazard logic determines whether the memory access type on the memory address range is subject to memory access rules contained in a rules table. If so, the memory hazard logic determines whether the memory access type matches a memory access rule associated with each memory address in the memory address range. For each such memory address, the memory hazard logic notifies the device regarding a memory hazard triggered by the memory access transaction.

Various embodiments discussed herein relate to generating a customizable route along a road network. Some embodiments generate a representation of a road network that includes locations and connections between those locations. Some embodiments generate a frequency-based lookup table that includes probabilities that a particular route can be selected from possible routes from one location to another. In response to receiving a request to generate a route, some embodiments select a location in the road network and, using the frequency-based lookup table, generate a route to another location in the road network. The route to the other location is then presented using a map interface.

This document relates to computer networking. For instance, the disclosed techniques provide for prioritizing packets based on Quality of Service ("QoS") levels within a networking device, such as a router. A router or other networking device consistent with the disclosed implementations can perform priority masking of received packets by comparing QoS priority values of the received packets to a designated QoS priority level. Packets having QoS priority values exceeding the designated QoS priority level can proceed to arbitration, whereas other packets can be masked to prevent them from proceeding to arbitration.

Some embodiments determine digital content to compose a software development environment and generate an executable image accordingly. In examples this includes obtaining a code artifact identification of a software repository, commit level, project, file extension, library, package, source code pattern, textual description, or other code artifact, and getting development environment component identifications which identify respective components, such as a software development tool, tool setting, tool extension, tool extension setting, security key or token or secret, runtime, kernel, driver, shell, or environment variable. These embodiments acquire correlation values that are calculated from historic data about artifact-component relationships. In response to results of comparing correlation values to thresholds, embodiments place components in a deployment set as default components of the environment being composed. Then these embodiments generate a deployable development environment executable image from at least the deployment set.

A chat history between a user and a machine learning model is preserved despite the context window size constraint of the machine learning model to ensure an enduring understanding of the chat or conversation history. When the token constraint of the context window size is reached, a summary of the chat history is generated to replace the chat history. The original content of the summarized chat history is stored in a lookup table and indexed by keywords. Instructions are provided to the model that allow the model to ask for the original content of the chat history summary which is obtained from the lookup table and provided to the model.

A display positioning assembly comprises a base portion and a display portion. The base portion comprises a base ferromagnetic component moveably retained within an elongated pocket of the base portion. The base ferromagnetic component is moveably retained within the elongated pocket. The display portion comprises a display ferromagnetic component. The display portion is moveably coupled to the base portion. One or both of the base ferromagnetic component and display ferromagnetic component is a magnet. The base portion further includes a trackpad, and at least a portion of the elongated pocket is disposed beneath the trackpad.

A dynamic reauthorization threshold is disclosed. A cloud customer is assigned a first reauthorization threshold associated with a first resource usage quota. When a current usage of the customer exceeds the first reauthorization threshold, a customer reauthorization process is performed. If the customer fails the reauthorization process, resources allocated to the customer are deallocated. If the customer passes the reauthorization process, the customer is assigned a next reauthorization threshold associated with a second resource usage quota. The next reauthorization threshold is determined based on at least one of: a second reauthorization threshold determined based at least on a variance of the current resource usage from the first resource usage quota, a base scaling factor, and a randomization factor, or a third reauthorization threshold determined by a machine learning model based at least on the first reauthorization threshold, the current resource usage, and the second reauthorization threshold.

According to examples, an adapter apparatus that enables the transformation of messages, based on storage hierarchies of different cloud storage systems in a heterogenous 5 cloud environment is disclosed. The adapter apparatus is included in a native cloud storage system and transforms messages from an external cloud storage system for execution on the native cloud storage system based on the native storage hierarchy. Similarly, the messages originating at the native cloud storage system are transformed for execution on the external cloud storage system based on the external storage hierarchy. A set of stored adapters 10 including a user-defined adapter, a catalog-based adapter and a model-based adapter enable the conversion. Various virtualization functions including cache control, virtualization of properties, and leases are also enabled by the transformations.

A method for mutual authentication between a host controller and a battery subsystem of an electronic device includes, at a battery controller, generating a battery-side authentication challenge, and transmitting the battery-side authentication challenge to the host controller. A host-side authentication challenge is received from the host controller. The host-side authentication challenge is signed with a private encryption key of the battery controller to generate a battery-side message digest. The battery-side message digest and a public encryption key of the battery controller are transmitted to the host controller as a battery-side signature pair. A host-side signature pair is received from the host controller. Based at least in part on the host-side signature pair, a battery-side validation result is output that specifies a battery-side validity state for the host controller.

Techniques are described herein that are capable of performing a security action based on a communication-based analysis. A security event, which is triggered by an operation performed by a user in an organization, is detected. A security analysis result is generated by determining whether a communication history of the user includes (1) a communication from the user that initiates an interaction with another user in the organization and/or a communication that is addressed specifically to the user from another user in the organization, (2) a communication from the user that references the operation, and/or (3) a communication that provides an explanation of a purpose of the operation that satisfies an explanation criterion. In response to the security event, a security action is performed with regard to the operation as a result of the security analysis result satisfying a security criterion.

A method, computer program product, and computing system for processing a natural language query concerning data stored within an electronic database. The natural language query is converted into a database command query using a generative AI model. A flagged ambiguity is generated by flagging an ambiguity associated with the natural language query using the generative AI model. User feedback concerning the flagged ambiguity is obtained. The database command query is revised by processing the natural language query, the database command query, and the flagged ambiguity using the generative AI model.

Context-aware content filtering adapted for a knowledge domain is provided. In certain examples, a classification confidence score by a classifier indicates a level of confidence that a prompt from a user is associated with the knowledge domain. The classification confidence score is compared with a threshold. When the score is below a threshold, a violation notice is provided to the user without submitting the prompt to a generative artificial intelligence (GAI) model. When the classification confidence score is above the threshold, the prompt is further processed to determine, according to rules, whether the prompt should be submitted to the GAI model. In various examples, the rules are applied to contextual information, safety score information, and intent information derived from the prompt.

A computer monitoring system may predict near term and long term performance by comparing a segment of current time series data with previously observed time series to find matching segments. From a matching segment, a prediction of performance may be made by examining later observations in the time series. Each time series element may include a large number of parameters, and one mechanism for comparing segments may be treating the elements as multidimensional vectors and using cosine similarity for finding significant matches. A deployment mechanism may store time series segments in a searchable database, and search the database with a newly observed time series segment for matches.

An exemplary implementation verifies an answer generated by a generative artificial intelligence model. The answer is responsive to an augmented query. The augmented query comprises a query submitted by a user and a request to answer the query using evidence determined to be relevant to the query. The answer is verified to be relevant to both the query and the evidence, facts are extracted from the evidence, and claims are extracted from the answer. Claim-to-fact entailment scores corresponding to each of the respective claim-to-fact pairings are determined. The scores are determined by an agreement analyzer comprising a natural language inference model. The answer is verified based on a check of the scores. A verified answer, a qualified answer, and/or a failure message is communicated to the user based on the verification.

Access management in a network-based service involves identifying an action involving a first account and determining if an account policy forbids the action. If forbidden, the system identifies an identity data structure linking the first account with a second account of a different type. The system determines if the second account allows the action, sends a transmission to the user's device identifying the second account, and receives an access token. If the token is valid, the system performs the action. The system may also create a new account if policies allow.

Access management in a network-based service involves identifying an action involving a first account and determining if an account policy forbids the action. If forbidden, the system identifies an identity data structure linking the first account with a second account of a different type. The system determines if the second account allows the action, sends a transmission to the user's device identifying the second account, and receives an access token. If the token is valid, the system performs the action. The system may also create a new account if policies allow.