Abstract: Techniques and tools for performing fading compensation in video processing applications are described. For example, during encoding, a video encoder performs fading compensation using fading parameters comprising a scaling parameter and a shifting parameter on one or more reference images. During decoding, a video decoder performs corresponding fading compensation on the one or more reference images.

Abstract: Various implementations relate to asynchronous training of a machine learning model. A server receives feedback data generated by training the machine learning model from a worker. The feedback data are obtained by the worker with its own training data and are associated with previous values of a set of parameters of the machine learning model at the worker. The server determines differences between the previous values and current values of the set of parameters at the server. The current value may have been updated for once or more due to operation of other workers. Then, the server can update the current values of the set of parameters based on the feedback data and the differences between values of the set of parameters. Thus, the updating does not only take the training result of each worker into consideration but also makes proper compensation for delay between different workers.

Abstract: The disclosed embodiments provide a system for processing data. During operation, the system obtains a set of coefficients from a linear model that uses a set of features inputted into a statistical model to estimate an output of the statistical model. Next, the system combines the set of coefficients with a set of feature values of the features to calculate a set of local contributions of the features toward the output of the statistical model, wherein each local contribution in the set of local contribution is calculated by multiplying each feature value in the set of feature values by a coefficient for a corresponding feature in the linear model. The system then outputs, based on a first ranking of the set of features by the set of local contributions, a first subset of the features for use in characterizing a local performance of the statistical model.

Abstract: Performance anomalies in production applications can be analyzed to determine the dynamic behavior over time of hosting processes on the same or different computers. Problematic call sites (call sites that are performance bottlenecks or that are causing hangs) can be identified. Instead of relying on static code analysis and development phase load testing to identify a performance bottleneck or application hang, a lightweight sampling strategy collects predicates representing key performance data in production scenarios. Performance predicates provide information about the subject (e.g., what the performance issue is, what caused the performance issue, etc.). The data can be fed into a model based on a decision tree to identify critical threads running the problematic call sites. The results along with the key performance data can be used to build a call graph prefix binary tree for analyzing call stack patterns. Data collection, analysis and visualizations of result can be performed.

Abstract: The disclosed embodiments provide a system for processing data. During operation, the system obtains a set of addresses for a set of entities. Next, for each address in the set of addresses, the system combines a set of verification rules and user input to generate a confidence in the address for a corresponding entity. The system then performs one or more steps for confirming the address according to the confidence in the address. Upon completing the one or more steps for confirming the address, the system stores the address for use with the corresponding entity.

Abstract: A method of and system for presentencing a user interface for the visually impaired is carried by dividing a user interface area of a user interface into a plurality of sections, mapping a different audible tone to each one of the plurality of sections, associating each one of the plurality of sections with a category, displaying a pane in one section of the plurality of sections, the pane having information about the category associated with the one section, receiving a first touch gesture on the pane, outputting a first audio signal having an audible tone mapped to the section, and depending on a type of the first touch gesture, outputting speech data for at least some of the information, the some of the information depending on a duration and a direction of the first touch gesture.

Abstract: Distribution and management of services in virtual environments is described herein. In one or more implementations, a service distribution and management model is implemented in which system services and applications are seamlessly distributed across multiple containers which each implement a different runtime environment. In one or more implementations, a system for distributing access to services in a host operating system of a computing device includes a host operating system configured to implement a host runtime environment, and one or more services implemented by the host operating system. The system further includes a service control manager configured to enable communication between a client stub of a service implemented in a client runtime environment and a service provider of the service that is implemented in a service runtime environment that is separate from the first client runtime environment.

Abstract: Disclosed herein is a system for providing information related to a shared frequency spectrum to a wireless network that lacks Internet connectivity. The system may include one or more processors and memory including instructions that are executable by the one or more processors to perform operations. The operations may include receiving registration of the wireless network when a vehicle is within a communication range of the wireless network, querying a database on behalf of the wireless network when an Internet connection is available, receiving query results including the information related to the shared frequency spectrum, and communicating the information related to the shared frequency spectrum to the wireless network when the vehicle or a proxy of the vehicle is again within the communication range of the wireless network.

Abstract: Aspects of the technology described herein can facilitate computing on transient resources. An exemplary computing device may use a task scheduler to access information of a computational task and instability information of a transient resource. Moreover, the task scheduler can schedule the computational task to use the transient resource based at least in part on the rate of data size reduction of the computational task. Further, a checkpointing scheduler in the exemplary computing device can determine a checkpointing plan for the computational task based at least in part on a recomputation cost associated with the instability information of the transient resource. Resultantly, the overall utilization rate of computing resources is improved by effectively utilizing transient resources.

Abstract: Aspects extend to methods, systems, and computer program products for changing between different roles at acceleration components. Changing roles at an acceleration component can be facilitated without loading an image file to configure or partially reconfigure the acceleration component. At configuration time, an acceleration component can be configured with a framework and a plurality of selectable roles. The framework also provides a mechanism for loading different selectable roles for execution at the acceleration component (e.g., the framework can include a superset of instructions for providing any of a plurality of different roles). The framework can receive requests for specified roles from other components and switch to a subset of instructions for the specified roles. Switching between subsets of instructions at an acceleration component is a lower overhead operation relative to reconfiguring or partially reconfiguring an acceleration component by loading an image file.

Abstract: A data processing stage is described which has a communications interface arranged to receive a plurality of input data chunks. Each input data chunk has a pointer to a validation record, where the validation records are stored at a memory accessible to the data processing stage. A processor of the data processing stage is configured to create an output validation record at the memory, and to link the output validation record to the validation records of the input chunks. The processor is configured to compute an output chunk from the input chunks in a manner which ignores data of the input chunks identified as invalid through inspection of the output validation record.

Abstract: Superconducting integrated circuits with clock signals distributed via an inductive coupling and related methods are provided. A method includes providing a D flip-flop having a clock terminal coupled to receive clock pulses from a clock line, a data input terminal, and a data output terminal. The D flip-flop may further include a first Josephson junction (JJ) coupled between a first terminal and a second terminal. The D flip-flop may further include a superconducting quantum interference device (SQUID) coupled between a third terminal and a fourth terminal, where an inductive loop, formed between the first JJ and the SQUID, is configured to in response to receiving a first clock pulse, store a fluxon when a state of the input data signal is high, and is configured to in response to receiving a second clock pulse to annihilate the stored fluxon when a state of the input data signal is low.

Abstract: Techniques for generating recommendation cluster within a social network service are described. Consistent with some embodiments, sample members in a social network service are identified. The sample members may be associated with prior member activity involving a source member. A cluster category this then selected based on a member attribute shared by a plurality of the sample members. In turn, a recommendation cluster is generated based on the selected cluster category. Generating the recommendation duster may involve selecting member profiles that match the cluster category. The member profiles selected in this way form the recommendation cluster. One or more of the member profiles of the recommendation cluster are then surfaced to a client device operated by the source member.

Abstract: Information management systems with time zone information, including event scheduling processes are disclosed. One aspect of the invention is directed toward a computer-implemented scheduling method that can include identifying a difference between a participant time zone and a user time zone, reviewing availability information for the participant and/or one or more selected time preference periods for the participant, and selecting a time range for an event. The method can further include reviewing one or more selected time preference periods for the user. Another aspect of the invention is directed toward a computer-implemented method for associating time zone information with a contact in an information management program application that includes selecting a contact, analyzing contact information associated with the contact, and determining a time zone for the contact based on the analysis of the contact information.

Abstract: Disclosed is an infrastructure to support an augmented reality (AR) based community for drivers. Each driver's view (e.g., using an AR device) of other vehicles on the road can be augmented with virtual information about other drivers in the community. Drivers in the community can express their opinions on the driving behavior of other drivers in the community. Drivers' opinions can be accumulated and tallied to produce a cumulative assessment of a particular behavior for a particular driver. A driver's view through their AR device and be further augmented with the community's assessment of that driver's driving behavior, which can influence the driver's future behavior.

Abstract: The techniques and systems described herein are directed to incremental sorting in a stream processing engine. For example, because of network delays, streaming data can be received at a stream processing engine that is out-of-order with respect to the timestamp of the event generation. Streaming data can be incrementally sorted by partitioning events into ordered runs, which provide intermediate arrays of ordered elements. Upon an assertion of a punctuation event (e.g., a progress indicator), partial runs can be output and merged to form sorted streaming data. Further, the sorted streaming data can be used to provide speculative outputs based on speculation latencies, which improves throughput and completeness of processing on the streaming data.

Abstract: The present disclosure provides methods, systems, and computer products for providing state-based revisions in a document collaboration system. In some aspects, a plurality of users may be simultaneously editing the same document on their respective computing devices. The document may comprise a plurality of editable objects having unique objects identifiers. In such aspects, when a user edits an object in a local version of the document, a state-based revision (“SBR”) record is created. The SBR is populated with information such as, information describing the type of object being edited, a descriptor that uniquely identifies the specific instance of the object being edited, a collection of common references that can be adjusted by the document collaboration engine and an opaque binary stream comprising the modified state of the edited object. The SBR record is then transmitted to one or more additional devices.

Abstract: An extension point virtualization system uses operating system-provided reparse points to provide minimal extension point registration. Reparse points preserve application isolation while removing the scale problem of writing custom extension point proxies for each extension point. Instead, the system can use a single file system filter that generically handles application virtualization reparse points, and store reparse point data for extension points that need redirection. Many extension points can be handled by redirecting the operating system from a typical location for an application resource to a virtualized safe location for the application resource. Thus, the system simplifies the process of handling new extension points by allowing an application virtualization system to simply register new locations that should be handled with reparse points and to then handle registered locations generically.

Abstract: Input context for a statistical dialog manager may be provided. Upon receiving a spoken query from a user, the query may be categorized according to at least one context clue. The spoken query may then be converted to text according to a statistical dialog manager associated with the category of the query and a response to the spoken query may be provided to the user.

Abstract: In one or more implementations, a static geometry model is generated, from one or more images of a physical environment captured using a camera, using one or more static objects to model corresponding one or more objects in the physical environment. Interaction of a dynamic object with at least one of the static objects is identified by analyzing at least one image and a gesture is recognized from the identified interaction of the dynamic object with the at least one of the static objects to initiate an operation of the computing device.