Abstract: A method of manufacturing an apparatus is provided. The apparatus is formed on a wafer or a package. The apparatus includes a polynomial generator, a plurality of matrix generators connected to an output of the polynomial generator, and a convolution generator connected to an output of the plurality of matrix generators. The apparatus is tested using one or more electrical to optical converters, one or more optical splitters, and one or more optical to electrical converters.

Abstract: A change in the emotional state of a user of a device is detected, and in response to this change a reason for the change in user emotional state is determined. This determination is made based on both the current user emotional state and context data for the device or user. The device then adapts to the user based on the current emotional state and the reason for the change in user emotional state. This adaptation of the computing device refers to an alteration of the operation of the computing device with a goal of increasing the likelihood of the user being in a good emotional state (e.g., happy, relaxed) and reducing the likelihood of the user being in a bad emotional state (e.g., sad, angry).

Abstract: An electronic device has pair(s) of conductive plates that is coupled to a housing and is electrically isolated by a dielectric material. The pair(s) of conductive plates is positioned to change relative orientation in relation to mechanical force that deforms the housing. Capacitive driver(s) is electrically coupled to pair(s) of conductive plates. A controller is communicatively coupled to the capacitive driver(s). The controller identifies a first capacitance of the pair(s) of conductive plates associated with the housing being in a normal state. The controller detects, via the capacitor driver(s), a change in capacitance from the first capacitance to a second capacitance of the pair(s) of conductive plates. The controller compares the change in capacitance to a threshold. In response to the change exceeding the threshold, the controller generates a notification.

Abstract: A client device configured with a neural network includes a processor, a memory, a user interface, a communications interface, a power supply and an input device, wherein the memory includes a trained neural network received from a server system that has trained and configured the neural network for the client device. A server system and a method of training a neural network are disclosed.

Abstract: Techniques for artificial intelligence (AI) modules for computation tasks are described, and may be implemented to enable multiple AI modules to be aggregated to form an execution structure (e.g., an execution chain) for performing a computation task. Generally, the described techniques aggregate AI modules based on their respective functions to perform an overall computation task.

Abstract: An electronic device includes a capacitive driver that is electrically coupled to a pair of conductive plates coupled to a housing and electrically isolated by a dielectric material. The conductive plates are positioned to change relative orientation in relation to deformation of the housing in a first direction. A controller identifies a first capacitance of the pair of conductive plates associated with the housing being in a normal alignment relative to the first direction. The controller presents an object on a user interface device in a first kinematic state. The controller detects, via the first capacitor driver, a change in capacitance from the first capacitance to a second capacitance of the pair of conductive plates. The controller associates the change to a first user input. The controller presents the object on the user interface device in a second kinematic state based on the first user input.

Abstract: A system and method to reduce weight storage bits for a deep-learning network includes a quantizing module and a cluster-number reduction module. The quantizing module quantizes neural weights of each quantization layer of the deep-learning network. The cluster-number reduction module reduces the predetermined number of clusters for a layer having a clustering error that is a minimum of the clustering errors of the plurality of quantization layers. The quantizing module requantizes the layer based on the reduced predetermined number of clusters for the layer and the cluster-number reduction module further determines another layer having a clustering error that is a minimum of the clustering errors of the plurality of quantized layers and reduces the predetermined number of clusters for the another layer until a recognition performance of the deep-learning network has been reduced by a predetermined threshold.

Abstract: A client device configured with a neural network includes a processor, a memory, a user interface, a communications interface, a power supply and an input device, wherein the memory includes a trained neural network received from a server system that has trained and configured the neural network for the client device. A server system and a method of training a neural network are disclosed.

Abstract: A client device configured with a neural network includes a processor, a memory, a user interface, a communications interface, a power supply and an input device, wherein the memory includes a trained neural network received from a server system that has trained and configured the neural network for the client device. A server system and a method of training a neural network are disclosed.

Abstract: Systems and methods are provided for generating a personalizable user interface (UI) component layout on a curved screen on a mobile electronic communication device, the curved screen having a front screen display surface and a plurality of edge screen display surfaces. In an example, the described technique entails detecting that a device user has opened an application on the mobile electronic communication device, determining whether the application opened by the user supports placement of UI components on edge screen display surfaces, and determining a coverage area on the front screen display surface and the edge screen display surfaces areas when the application supports placement of UI components on edge screen display surfaces, and estimating a user hand size based on the determined coverage area, and placing one or more UI components on an edge screen display surface to be reached by the user based on the estimated hand size.

Abstract: In aspects of generating action suggestions based on a change in user mood, a computing device implements an action generation module that generates action suggestions. The action generation module analyzes sensor data from sensors to determine a current user state associated with a user. The action generation module also analyzes environment data received from Internet-of-Things (IoT) devices to determine a current environment context associated with the user. The action generation module can detect a change in a user mood based on a comparison of the current user state with user history data that includes user state parameters and environment context parameters that indicate the user mood. The action generation module can analyze the user history data to determine the user state parameters and/or the environment context parameters that correlate to the change in the user mood, and generate an action suggestion that is intended to modify the current user state.

Abstract: Content visibility on a computing device is controlled based at least in part on the current location of the computing device. When the current location is a safe location for making user-selectable content visible on a display screen of the computing device and the computing device is unlocked, the computing device operates in a full content visibility mode. In the full content visibility mode all user-selectable content on the computing device is accessible to a user of the computing device. When the current location is not a safe location for making user-selectable content visible on the display screen, the computing device operates in a reduced content visibility mode. In the reduced content visibility mode content visibility on the display screen is reduced, such as by limiting which applications (e.g., application icons or widgets) are displayed or otherwise accessible to the user.

Abstract: A system and method to reduce weight storage bits for a deep-learning network includes a quantizing module and a cluster-number reduction module. The quantizing module quantizes neural weights of each quantization layer of the deep-learning network. The cluster-number reduction module reduces the predetermined number of clusters for a layer having a clustering error that is a minimum of the clustering errors of the plurality of quantization layers. The quantizing module requantizes the layer based on the reduced predetermined number of clusters for the layer and the cluster-number reduction module further determines another layer having a clustering error that is a minimum of the clustering errors of the plurality of quantized layers and reduces the predetermined number of clusters for the another layer until a recognition performance of the deep-learning network has been reduced by a predetermined threshold.

Abstract: A system and method to reduce weight storage bits for a deep-learning network includes a quantizing module and a cluster-number reduction module. The quantizing module quantizes neural weights of each quantization layer of the deep-learning network. The cluster-number reduction module reduces the predetermined number of clusters for a layer having a clustering error that is a minimum of the clustering errors of the plurality of quantization layers. The quantizing module requantizes the layer based on the reduced predetermined number of clusters for the layer and the cluster-number reduction module further determines another layer having a clustering error that is a minimum of the clustering errors of the plurality of quantized layers and reduces the predetermined number of clusters for the another layer until a recognition performance of the deep-learning network has been reduced by a predetermined threshold.

Abstract: An electronic device includes one or more low-power context sensors, one or more high-power context sensors, and one or more processors operable with the one or more low-power context sensors and the one or more high-power context sensors. The one or more processors, working with context engines associated with the sensors, minimize usage of the high-power context sensors when determining a context of the electronic device, where that determined context has a confidence score above a predefined confidence level threshold.

Abstract: Systems and methods for providing smart internet of things (IOT) menus utilize a camera of an electronic communications device and entail capturing an image of an IOT device of interest using a camera of the electronic communications device and identifying the IOT device of interest. The IOT of interest is displayed and may be selected by the device user. The user of the electronic communications device is identified and their historical behavior with respect to the selected IOT device is analyzed. A customized, rather than full, IOT device control menu is then generated and presented to the user.

Abstract: A system and method of adaptive interface switching in a mobile electronic communications device entails engaging in an audio interface task at the mobile electronic communications device. If during the audio interface task it is detected that at least one microphone or at least one speaker of the mobile electronic communications device is compromised, then the system activates a data link to a wearable communication device associated with a user of the mobile electronic communications device. At least one of a mic and speaker on the wearable communication device is then used to continue the audio interface task.

Abstract: An electronic device includes a capacitive driver that is electrically coupled to a pair of conductive plates coupled to a housing and electrically isolated by a dielectric material. The conductive plates are positioned to change relative orientation in relation to deformation of the housing in a first direction. A controller identifies a first capacitance of the pair of conductive plates associated with the housing being in a normal alignment relative to the first direction. The controller presents an object on a user interface device in a first kinematic state. The controller detects, via the first capacitor driver, a change in capacitance from the first capacitance to a second capacitance of the pair of conductive plates. The controller associates the change to a first user input. The controller presents the object on the user interface device in a second kinematic state based on the first user input.

Abstract: An electronic device has pair(s) of conductive plates that is coupled to a housing and is electrically isolated by a dielectric material. The pair(s) of conductive plates is positioned to change relative orientation in relation to mechanical force that deforms the housing. Capacitive driver(s) is electrically coupled to pair(s) of conductive plates. A controller is communicatively coupled to the capacitive driver(s). The controller identifies a first capacitance of the pair(s) of conductive plates associated with the housing being in a normal state. The controller detects, via the capacitor driver(s), a change in capacitance from the first capacitance to a second capacitance of the pair(s) of conductive plates. The controller compares the change in capacitance to a threshold. In response to the change exceeding the threshold, the controller generates a notification.

Abstract: Techniques for artificial intelligence (AI) modules for computation tasks are described, and may be implemented to enable multiple AI modules to be aggregated to form an execution structure (e.g., an execution chain) for performing a computation task. Generally, the described techniques aggregate AI modules based on their respective functions to perform an overall computation task.