Abstract:
Floating-point numbers are compressed for neural network computations. A compressor receives multiple operands, each operand having a floating-point representation of a sign bit, an exponent, and a fraction. The compressor re-orders the operands into a first sequence of consecutive sign bits, a second sequence of consecutive exponents, and a third sequence of consecutive fractions. The compressor then compresses the first sequence, the second sequence, and the third sequence to remove at least duplicate exponents. As a result, the compressor can losslessly generate a compressed data sequence.
Abstract:
A radio frequency (RF) communication assembly includes an RF communication circuit and a compensator apparatus. The compensator apparatus receives an input including an I-component of a pre-compensated signal, a Q-component of the pre-compensated signal, and encoded operating conditions of the RF communication circuit. The RF communication circuit includes RF circuit components causing signal impairments. The compensator apparatus perform neural network computing on the input, and the RF communication assembly generates a compensated output signal that compensates for at least a portion of the signal impairments.
Abstract:
A method for generating a user scenario of an electronic device includes detecting a real part and an imaginary part of an input impedance of each antenna of the electronic device, using a plurality of sensors of the electronic device to generate a plurality of sensing signals, and entering at least the real part and the imaginary part of the input impedance of each antenna, and the plurality of sensing signals to a machine learning model to output the user scenario.
Abstract:
An electronic device includes an application processor and a sensor hub, where the application processor is arranged for executing applications running on a system of the electronic device, and the sensor hub is arranged for obtaining and processing sensed data from a plurality of sensors within the electronic device. In addition, the application processor further downloads location data from a remote device via a network module, and at least a portion of the downloaded location data is further stored in a storage unit of the sensor hub to be reused for positioning.
Abstract:
A dynamic data distribution method in a private network and an associated electronic device are provided. The private network includes: a first pairing connection between a first electronic device, a second electronic device, and a second pairing connection between the first electronic device and a third electronic device. The method includes the steps of: receiving sensor data from the second electronic device by the first electronic device; notifying the second electronic device to build a third pairing connection with the third electronic device according to a determination result between the first electronic device and the third electronic device; and terminating the first pairing connection and retrieving the sensor data from the second electronic device through the third electronic device by the first electronic device when the third pairing connection has been built.
Abstract:
A method for accessing a network in an electronic system and associated portable device are provided. The portable device includes; a transceiver, supporting a plurality of predetermined communication protocols; and a processor, configured to connect the portable device to a connectivity service device in an electronic system via the transceiver when the portable device enters a coverage region of the connectivity service device. The connectivity service device retrieves service information from a plurality of electronic devices that are connected to the connectivity service device, to build a service list. The processor retrieves the service list from the connectivity service device, and determines a service from the service list to be used for communicating with the plurality of the electronic devices.
Abstract:
A radio frequency (RF) communication assembly includes an RF communication circuit and a compensator apparatus. The compensator apparatus receives an input including an I-component of a pre-compensated signal, a Q-component of the pre-compensated signal, and encoded operating conditions of the RF communication circuit. The RF communication circuit includes RF circuit components causing signal impairments. The compensator apparatus perform neural network computing on the input, and the RF communication assembly generates a compensated output signal that compensates for at least a portion of the signal impairments.
Abstract:
The present invention provides a data aggregator serving between at least one data source and at least one electronic device, wherein the data aggregator is arranged to wirelessly communicate with the data source and the electronic device, and the data aggregator comprises a memory comprising a data cache, a backup memory and a FIFO buffer, and a controller for controlling a use of the memory. The controller selects at least one of the data cache, the backup memory and the FIFO buffer to store data received from the data source according to characteristics of the data provided by the data source, and forwards the data stored in the data cache, the backup memory or the FIFO buffer to the electronic device.
Abstract:
A wireless device includes a satellite receiver to receive data from multiple satellites. The wireless device also includes processing circuitry and memory. The memory stores one or more neural network models. The processing circuitry is operative to identify a neural network model that has been trained to adapt to a region in which the wireless device operates, classify satellite raw measurements from each satellite at a given time into a corresponding quality level using the neural network model, and identify satellite raw measurements with a quality level higher than a threshold. The location of the wireless device is calculated using the identified satellite raw measurements.
Abstract:
An arbitrary-scale blind super resolution model has two designs. First, learn dual degradation representations where the implicit and explicit representations of degradation are sequentially extracted from the input low resolution image. Second, process both upsampling and downsampling at the same time, where the implicit and explicit degradation representations are utilized respectively, in order to enable cycle-consistency and train the arbitrary-scale blind super resolution model.