Abstract: Systems and methods for object detection generate a feature pyramid corresponding to image data, and rescaling the feature pyramid to a scale corresponding to a median level of the feature pyramid, wherein the rescaled feature pyramid is a four-dimensional (4D) tensor. The 4D tensor is reshaped into a three-dimensional (3D) tensor having individual perspectives including scale features, spatial features, and task features corresponding to different dimensions of the 3D tensor. The 3D tensor is used with a plurality of attention layers to update a plurality of feature maps associated with the image data. Object detection is performed on the image data using the updated plurality of feature maps.

Abstract: A system for generating a user interface described herein can include a processor to detect a plurality of display characteristics from a user interface manager, wherein the plurality of display characteristics correspond to a type of a device. The processor can also detect a list of applications being executed by the system and generate a user interface container by applying the plurality of display characteristics to each of the applications from the list of applications.

Abstract: Methods and apparatus for selectively performing speech processing in a hybrid speech processing system. The hybrid speech processing system includes at least one mobile electronic device and a network-connected server remotely located from the at least one mobile electronic device. The mobile electronic device is configured to use an embedded speech recognizer to process at least a portion of input audio to produce recognized text. A controller on the mobile electronic device determines whether to send information from the mobile electronic device to the server for speech processing. The determination of whether to send the information is based, at least in part, on an analysis of the input audio, the recognized text, or a semantic category associated with the recognized text.

Abstract: According to examples, an apparatus may include a processor that may, responsive to a secondary display being detected, cause a window including an option for a user to proceed with a location detection operation of the secondary display to be displayed on a primary display. The processor may also cause an icon to be displayed on the secondary display, track a movement of a cursor from the primary display to the icon displayed on the secondary display, and determine a position of the secondary display with respect to the primary display based on the tracked movement of the cursor. The processor may further store the determined position of the secondary display with respect to the primary display.

Abstract: Described herein are supervised machine learning techniques for generating a first machine learned model, which receives, as input, values for a set of utilization metrics for a server computer and generates, as output, a service-specific health score for the server computer, and for generating a second machine learned model, which receives, as input, values for the volume of application traffic directed to a group of server computers hosting the service for each of several prior time periods and generates, as output, an application traffic forecast for the group of server computers, for a future time period.

Abstract: Logging cache line lifetime hints when recording an execution trace. A microprocessor detects occurrence of a first cache event that initiates a lifetime of a cache line within a memory cache, and initiates logging first trace information indicating a beginning of the lifetime of the cache line within the memory cache. Subsequently, the microprocessor detects occurrence of a second cache event that ends the lifetime of the cache line within the memory cache. Based on detecting the second cache event, the microprocessor initiates logging second trace information indicating an ending of the lifetime of the cache line within the memory cache.

Abstract: Email has become commonplace as a means to communicate among parties non-synchronously by exchanging email messages. Some communications, however, may be more effective in real-time or in near real-time. For instance, service providers may offer a chat feature to accommodate real-time or near real-time help or service communications with customers. However, in this case, the user is likely required to enter credentials and/or context information to transition into a different communication type with the service provider. The present systems and methods provide a continuous conversation experience for the user whereby a message recipient of one message type (e.g., email, text, etc.) may continue a conversation in a different message type (e.g., a chat session, instant message, etc.) without manually providing message context for communicating via the second message type.

Abstract: Examples are disclosed that relate to improving speed and accuracy of touch input classification. In one example, a touch detection device includes an array of antennas configured to measure touch input and output a touch matrix of pixels having touch values corresponding to the touch input measured at each antenna of the array of antennas. The touch detection device further includes a neural network having an input layer including a plurality of nodes. Each node is configured to receive a touch value corresponding to a different pixel of the touch matrix. The neural network is configured to output classified touch data corresponding to the measured touch input based at least on the touch matrix.

Abstract: Methods, systems, and apparatuses are provided for detecting a missing security alert by receiving an alert sequence generated by a network security provider, applying the received alert sequence to a security incident model, receiving an indication from the security incident model that the received alert sequence corresponds to a security incident defined by a predetermined sequence of alerts that includes at least one alert missing from the received alert sequence, and generating a notification to the network security provider that indicates at least one of the security incident or the missing alert(s). In addition, the security incident model may be generated by providing a set of historical alerts and a set of historical security incidents to a machine learning algorithm to generate the security incident model.

Abstract: A computing device includes a system-on-a-chip. The computing device comprises a network interface controller (NIC) that hosts a plurality of virtual functions and physical functions. Two or more compute nodes are coupled to the NIC. Each compute node is configured to operate a plurality of Virtual Machines (VMs). Each VM is configured to operate in conjunction with a virtual function via a virtual function driver. A dedicated VM operates in conjunction with a virtual NIC using a physical function hosted by the NIC via a physical function driver hosted by the compute node. The computing device further comprises a fabric manager configured to own a physical function of the NIC, to bind virtual functions hosted by the NIC to individual compute nodes, and to pool I/O devices across the two or more compute nodes.

Abstract: Disclosed in some examples are methods, systems, devices, and machine-readable mediums which provide for sidebar communication threads forked from, or related to, a principal thread. Messages in the sidebar communication thread may include a history of the principal thread, including one or more messages from the principal thread, and may include a proposed principal thread message that is the subject of the sidebar thread discussion. The sidebar thread may also include sidebar thread messages that carries the conversation of the sidebar thread participants. Once a termination condition is reached for the sidebar thread, the sidebar thread terminates and either the proposed principal thread message (as potentially modified by participants of the sidebar thread) becomes an accepted principal thread message and it is posted to the principal thread as if it was sent by the sidebar initiator or no message is posted (e.g., the proposed principal thread message is rejected).

Abstract: The performance of a text parser implemented with a state machine is improved by reducing a critical dependence path. In one aspect, all possible current states for a given text input are read from a state table circuit, and the correct next state and output are then selected therefrom by an output multiplexer based on the current state, removing dependence on the current state from the table read, and allowing the read(s) to be pipelined. Further, multiple input units are configured to operate on multiple text characters in parallel, with each input unit propagating outputs for its state table circuit to the next downstream input unit. Each downstream input unit is configured to use the propagated states to provide the proper outputs to appropriates multiplexer inputs. The number of possible output states may be dynamically reduced, thereby reducing the size of the output multiplexer needed to select the next state.

Abstract: Techniques of deploying fuel cells in a facility are described herein. In one embodiment, a method includes identifying a location of the receptacle at the facility that the fuel cell is connected upon detecting the fuel connector of the second side of the carrier being coupled to a fuel port at a receptacle at the facility. The method can then include generating and storing, in a database, a fuel cell record indicating that the fuel cell is physically connected to the receptacle at the identified location in the facility and instructing a control device in the facility corresponding to the identified location to provide fuel to the fuel cell via the fuel port, the fuel connector, the connection between the first side and the second side of the carrier, and the fuel inlet of the fuel cell.

Abstract: A particular method includes storing a manifest file and a parameters file at a data store. The manifest file and the parameters file are associated with a software deployment package. An application associated with the software deployment package is published to a first remote computing device and to a second remote computing device. The method includes receiving first parameter values from the first remote computing device and receiving second parameter values from the second remote computing device. The first parameter values are stored in a first region of the data store that is specific to the first computing device. The second parameter values are stored in a second region of the data store that is specific to the second computing device.

Abstract: An electrical connector includes a receptacle connector having a tongue member with a surface extending between first and second tongue ends and having a shoulder extending from the surface adjacent to the second tongue end. The shoulder has a cross-sectional area within a receptacle opening size limit. The receptacle connector further includes a set of electrical contacts each extending through the shoulder between first and second contact ends on opposite sides of the shoulder. The electrical contacts at the first contact end are attached to the tongue surface at the first tongue end. The electrical contacts at the second contact end extend beyond the second tongue end. The receptacle connector further includes an electromagnetic interference (EMI) shielding bracket covering at least a portion of the set of electrical contacts between the first contact end and the shoulder. The disclosure further includes a method of making the electrical connector.

Abstract: Handling a memory fault based on detecting whether a memory pointer was invalidated by a pointer authentication (PA) failure. After an access to a memory pointer causes a memory fault, detecting that the memory pointer was invalidated by a PA failure includes creating a new memory pointer by replacing reserved bits of the memory pointer with a default value, and determining that the new memory pointer corresponds to a memory address that falls within executable memory. This determination includes determining that the memory address is within an executable memory page, determining that a call instruction is stored at a prior memory address that immediately precedes the memory address, and/or determining that the memory address corresponds to a code section of an executable file. The PA failure is handled based on logging the PA failure, terminating the application program, and/or resuming execution at an instruction stored at the memory address.

Abstract: The discussion relates to cooling computing devices and specifically to managing two-phase cooling. One example can include a two-phase liquid immersion tank containing heat generating components and a liquid phase of a coolant having a boiling point within an operating temperature range of the heat generating components. The example can also include a stratification chamber fluidly coupled to the liquid immersion tank and configured to at least partially separate a gas phase of the coolant from other gases. The example can further include a condenser chamber fluidly coupled to the stratification chamber and configured to receive the gas phase of the coolant and cause the gas phase of the coolant to phase change back into the liquid phase of the coolant.

Abstract: A system for power efficient image acquisition is configurable to capture, using an image sensor, a plurality of partial image frames including at least a first partial image frame and a second partial image frame. The first partial image frame is captured at a first timepoint using a first subset of image sensing pixels of the plurality of image sensing pixels of the image sensor. The second partial image frame is captured at a second timepoint using a second subset of image sensing pixels of the plurality of image sensing pixels of the image sensor. The second subset of image sensing pixels includes different image sensing pixels than the first subset of image sensing pixels, and the second timepoint is temporally subsequent to the first timepoint. The system is configurable to generate a composite image frame based on the plurality of partial image frames.

Abstract: A transcription of audio speech included in electronic content associated with a meeting is created by an ASR model trained on speech-to-text data. The transcription is post-processed by modifying text included in the transcription, for example, by modifying punctuation, grammar, or formatting introduced by the ASR model and by changing or omitting one or more words that were included in both the audio speech and the transcription. After the transcription is post-processed, output based on the post-processed transcription is generated in the form of a meeting summary and/or template.

Abstract: In an example embodiment, a deep learning algorithm is introduced that operates on a transition matrix formed from user activities in an online network. The transition matrix records the frequencies that particular transitions between paths of user activity have occurred (e.g., the user performed a login activity, which has one path in the network, and then performed a profile view action, which has another path in the network). Each transition matrix corresponds to a different user's activities.