Abstract: This document discloses aspects of secure serial peripheral interface (SPI) communication. In some aspects, a secure SPI communication module monitors communications transmitted by a host to a peripheral block that is coupled to the host via a SPI interconnect. The module compares respective commands of the communications sent by the host to information indicating commands that the peripheral block is not authorized to execute. Based on the comparing, the module determines that one of the respective commands is one of the commands that the peripheral block is not authorized to execute. The module then prevents the peripheral block from receiving at least a portion of the respective command of the communication. By so doing, the module can prevent the peripheral block from executing unauthorized commands, which may compromise security of the peripheral block.

Abstract: A method includes receiving a sequence of acoustic frames characterizing an input utterance and generating a higher order feature representation for a corresponding acoustic frame in the sequence of acoustic frames by an audio encoder of an automatic speech recognition (ASR) model. The method also includes generating a context embedding corresponding to one or more previous transcriptions output by the ASR model by a context encoder of the ASR model and generating, by a prediction network of the ASR model, a dense representation based on a sequence of non-blank symbols output by a final Softmax layer. The method also includes generating, by a joint network of the ASR model, a probability distribution over possible speech recognition hypotheses based on the context embeddings generated by the context encoder, the higher order feature representation generated by the audio encoder, and the dense representation generated by the prediction network.

Abstract: Active noise cancellation systems and methods include a feedforward path configured to receive a reference signal comprising ambient noise and adaptively generate an anti-noise signal to cancel the ambient noise. The adaptive filter is tuned in accordance with at least one parameter, which is set by a logic device configured to determine an ambient noise condition based the reference signal by estimating a fullband power of the reference signal, estimating a low-frequency power of the reference signal, comparing the fullband power and low-frequency power to one or more thresholds, and/or setting one or more ambient noise flags. The ambient noise condition may include a quiet background, a wideband noise condition, and/or a low-frequency dominant noise condition.

Abstract: A method for training a modular neural network model includes training only a backbone model to provide a first model configuration of the modular neural network model. The first model configuration includes only the trained backbone model. The method also includes adding an intrinsic sub-model to the trained backbone model. During a fine-tuning training stage, the method includes freezing parameters of the trained backbone model and fine-tuning parameters of the intrinsic sub-model added to the trained backbone model while the parameters of the trained backbone model are frozen to provide a second model configuration that includes the backbone model initially trained during the initial training stage and the intrinsic sub-model having the parameters fine-tuned during the fine-tuning stage.

Abstract: A firewall intelligence system, includes a data storage storing a set of firewall rules for a network; a recommendation engine that receives, from a log service, traffic logs detailing traffic for the network and firewall logs detailing the usage of firewall rules in response to the traffic for the network, accesses, from the data storage, the set of firewall rules for the network; processes the set of firewall rules to evaluate the firewall rules against a set of quantitative evaluation rules to determine one or more firewall rule recommendations, wherein each firewall rule recommendation is a recommendation to change at least one of the firewall rules in the set of firewall rules; and a front end API that provides data describing the one or more firewall rule recommendations to a user device.

Abstract: In a display assistant device, a speaker is mounted in a waveguide structure which is at least partially disposed beneath a display screen. The waveguide structure is mounted in an exterior housing which includes speaker grills distributed on a plurality of surfaces of the exterior housing, permitting sound waves from the speaker to be projected outside the exterior housing. A cover structure is disposed on top of the waveguide structure to conceal the waveguide structure and speaker within the exterior housing. The cover structure has a tilted bottom surface configured to be suspended above the waveguide structure and to be separated by a first space. Sound waves projected from an upper portion of the speaker are reflected by the tilted bottom surface and are guided through the first space to exit the device from a speaker grill portion located on a rear side of the exterior housing.

Abstract: An apparatus with an integrated circuit (IC) chip can provide protection against attacks on a read-only memory (ROM), such as a boot ROM for security circuitry. An attacker can gain control of an IC by modifying ROM contents and/or redirecting ROM reads. To combat these attacks, example implementations store encrypted ROM data in the ROM array. A ROM controller is used to cryptographically tie the address of each ROM line to the corresponding encrypted ROM datum. To access the encrypted ROM datum, cryptographic circuitry decrypts the encrypted ROM datum using a key that is generated based on the corresponding ROM address. As part of an integrity checking procedure, a digest can be computed based on the encrypted ROM data. To further thwart would-be attacks, the ROM address can be adjusted (e.g., scrambled) before the controller uses the adjusted address to read encrypted data from the ROM array.

Abstract: Various arrangements for integrating control of multiple cloud-based smart-home devices are presented. Registration information may be received for a first and second smart-home device that are controlled using different cloud-based server systems. A determination may be made that that the first smart-home device and the second smart-home device share a common function. The first smart-home device and the second smart-home device may be assigned to a common operating characteristic group based on the common function being shared by the first smart-home device and the second smart-home device. A control element may be provided that allows for control of smart-home devices with the common operating characteristic group. The control element may be used to control the common function.

Abstract: A method for adapting a guest mode for portable speakers includes obtaining a current state of an assistant-enabled device while a guest mode is enabled. Here, the assistant-enabled device is configured to allow a digital assistant to fulfill queries spoken by speakers who are not enrolled users when the guest mode is enabled. Based on the current state of the assistant-enabled device, the method includes disabling the guest mode to suppress fulfillment of queries spoken by speakers who are not enrolled users. While the guest mode is disabled, the method includes receiving audio data corresponding to a query, the query including a command for the digital assistant to perform an action, performing speaker identification to determine that a speaker of the query is not an enrolled user, and based on determining that the speaker of the query is not an enrolled user, suppressing fulfillment of the action.

Abstract: This document describes systems and techniques directed at autogenerated language model notifications. In aspects, a device management system associated with a network of network-connected devices includes a prompt manager. The prompt manager obtains metadata associated with the network and integrates at least portions of data from the metadata into one or more templated prompts to create an instantiated prompt. The instantiated prompt can be transmitted to a language model to generate a language model output. The prompt manager may then provide, based on the language model output, a notification to a user associated with at least one network-connected device. Through such a technique, the prompt manager can improve user experience and facilitate user proactiveness managing their network of network-connected devices.

Abstract: A method includes obtaining an ASR model trained to recognize speech in a first language and receiving transcribed training utterances in a second language. The method also includes integrating the ASR model with an input reprogramming module and a latent reprogramming module. The method also includes adapting the ASR model to learn how to recognize speech in the second language by training the input reprogramming module and the latent reprogramming module while parameters of the ASR model are frozen.

Abstract: A circuit for controlling an interface between an integrated circuit of an electronic device and an external device includes each of an interface control switch and an overvoltage protection switch. The interface control switch includes a controller for controlling a mode of the interface between a data communication mode that connects the external device to the integrated circuit of the electronic device via one or more communication lines of the electronic device, and a power transmission mode that connects a power source of the external device to a battery of the electronic device. The controller includes a reset line for resetting the interface control switch. The overvoltage protection switch is connected between the one or more communication ports and the reset line, and is configured to activate the reset line of the controller in response to an overvoltage condition at the one or more communication lines of the electronic device.

Abstract: A method of providing a personalized audio briefing to a user is performed at an electronic device. The device receives identification of information sources associated with the user. Each of the information sources is associated with a content type. The device receives an authorization to access the identified information sources and a preferred order of content types for presentation within the audio briefing. It transmits to a remote system the identification, the authorization, and the preferred order. Following the transmitting, the device receives a verbal input from the user requesting the audio briefing. In response to the verbal input, the device receives a response generated by the remote system, including content from the information sources and information inserted by the remote system based on the authorization to access received from the user. The device outputs an audible response according to the preferred order.

Abstract: Systems, methods, devices, and techniques for planning travel of an autonomous robot. A system identifies one or more obstacles that are located in proximity of at least a portion of a planned route for the autonomous robot. For each obstacle, the system: (i) determines a semantic class of the obstacle, including selecting the semantic class from a library that defines a set of multiple possible semantic classes for obstacles, and (ii) selects a planning policy for the obstacle that corresponds to the semantic class of the obstacle. The system can generate a trajectory along the at least the portion of the planned route using the selected planning policies. The robot can then initiate travel according to the trajectory.

Abstract: Techniques and apparatuses are described that implement an ultrasonic sensor capable of detecting user presence. This ultrasonic sensor can detect user presence without relying on time-of-flight techniques. In particular, the ultrasonic sensor can determine that a user is present based on the occlusion of at least one receiving transducer (e.g., microphone occlusion), the occlusion of at least one transmitting transducer (e.g., speaker occlusion), or a detected change in an audible noise floor of at least one transducer. In this way, the ultrasonic sensor can continue to detect user presence in situations in which a user occludes one or more transducers of the ultrasonic sensor. The ultrasonic sensor can also control operation of another component within a computing device based on the presence of the user to improve the user experience and/or improve power management.

Abstract: A server device is configured to receive criteria that indicate when a microsite is to be provided to a user device. The user device is configured to obtain sensor data and user profile data and send the sensor data and the user profile data to the server device. The server device is configured to receive the sensor data and the user profile data, identify a relevant microsite, and provide information regarding the relevant microsite to the user device. In some cases, the service device is configured to identify that a particular microsite has been purchased by a user of the user device and provide information, to the user device, identifying that the particular microsite has been purchased.

Abstract: Systems and methods for trigger event determination can include processing blockchain data to determine a trigger event. Data associated with the trigger event can be processed to determine a query. The query can then be utilized to recursively search a database for data descriptive of the trigger event occurring. A notification can then be provided to instruct a resulting action to be performed.

Abstract: Devices for wireless charging with metal housings. The device comprises a wireless power receiving coil and an associated coil core, wherein the coil core defines a recess and at least a portion of the wireless power receiving coil is disposed in the recess; and a housing comprising one or more metal portions, wherein the housing is configured to receive at least a portion of the wireless power receiving coil and the associated coil core in an opening defined by the one or more metal portions. Alternatively, the device comprises a wireless power receiving coil and an associated coil core; a housing comprising one or more metal portions; and an alignment feature configured to interact with a magnet of a wireless charger to align the device with the wireless charger, wherein the alignment feature comprises a plurality of metal elements that are electrically insulated from each other.

Abstract: Generally, the present disclosure is directed to systems and methods that perform adaptive optimization with improved convergence properties. The adaptive optimization techniques described herein are useful in various optimization scenarios, including, for example, training a machine-learned model such as, for example, a neural network. In particular, according to one aspect of the present disclosure, a system implementing the adaptive optimization technique can, over a plurality of iterations, employ an adaptive per coordinate clipping threshold to clip a current first moment of the coordinate to obtain a current update value that enables faster convergence for the machine-learned model when the noise in the stochastic gradients is heavy tailed.

Abstract: The present disclosure is directed to systems and methods for providing a stable topological representation of pathway networks as well as features associated with these networks. The disclosure is exemplified using road networks which have applications in mapping, navigation, and autonomous vehicles. Extensions may be learned through practice of the disclosure. Utilizing implementations disclosed herein may provide advantages for data conflation between different mapping systems and map data while improving overall stability by developing a common reference standard that is tied to semantic features rather than abstract geographic representations.