Abstract: This disclosure provides a voice interaction method and apparatus. At least one target device is determined from various electronic devices by receiving a wakeup signal sent by at least one wakeup device in the various electronic devices in response to a voice instruction of a user. Each target device is instructed to recognize a semantic meaning of the voice instruction to obtain semantic information, so as to determine a first response device in the various target devices based on the semantic information. In this case, the first response device is enabled to respond to the voice instruction. In this way, a problem that it is difficult for a plurality of devices to interact with the user accurately is resolved, thereby facilitating user experience.

Abstract: Disclosed are a method and an apparatus for performing an operation of a convolutional layer in a convolutional neural network.

Abstract: Disclosed is a method for convolution calculation in a neural network, comprising: reading an input feature map, depthwise convolution kernels and pointwise convolution kernels from a dynamitic random access memory (DRAM); performing depthwise convolution calculations and pointwise convolution calculations according to the input feature map, the depthwise convolution kernels and the pointwise convolution kernels to obtain output feature values of a first predetermined number p of points on all pointwise convolution output channels; storing the output feature values of a first predetermined number p of points on all pointwise convolution output channels into an on-chip memory, wherein the first predetermined number p is determined according to at least one of available space in the on-chip memory, a number of the depthwise convolution calculation units, and width, height and channel dimensions of the input feature map; and repeating the above operation obtain output feature values of all points on all pointwis

Abstract: A method and an apparatus for adapting feature data in a convolutional neural network. The method includes selecting a plurality of consecutive layers; determining an expected number of subdata blocks and a layout position, width and height of each subdata block in an output feature data of a last layer; determining, for each current layer, a layout position, width, and height of each subdata block of an input feature data for the current layer according to the layout position, width, and height of each subdata block of the output feature data for the current layer; determining an actual position of each subdata block of the input feature data for a first layer in the input feature data for the first layer; and obtaining the expected number of subdata blocks of the input feature data for the first layer according to the actual position, width and height of each subdata block of the input feature data for the first layer.

Abstract: Disclosed are a method and an apparatus for performing convolution operation on folded feature data. The method comprises: reading the folded feature data provided to a convolution layer and an original convolution kernel from a dynamic random access memory (DRAM); pre-processing the folded feature data and the original convolution kernel; storing the pre-processed folded feature data into a static random-access memory (SRAM); folding the pre-processed original convolution kernel in at least one dimension of width or height according to a folding manner of the folded feature data to generate one or more folded convolution kernels corresponding to the original convolution kernel; storing the one or more folded convolution kernels in the SRAM; and reading the pre-processed folded feature data and the one or more folded convolution kernels from the SRAM into a calculation unit for convolving the pre-processed folded feature data with the one or more folded convolution kernels.

Abstract: Disclosed are a method and an apparatus for performing an operation of a convolutional layer in a convolutional neural network. The method includes reading unfolded-feature-data and an original convolution kernel from DRAM, padding the unfolded-feature-data, folding the padded unfolded-feature-data in at least one dimension folded feature data, storing the folded feature data into a SRAM, folding the original convolution kernel in the at least one dimension to generate one or more folded convolution kernels, storing the one or more folded convolution kernels in the SRAM and reading the folded feature data and the one or more folded convolution kernels from the SRAM into a calculation circuit for performing a convolution operation on the folded feature data by using the one or more folded convolution kernels.

Abstract: Disclosed are a method and an apparatus for adapting parameters of a neural network. The method includes selecting one or more dimensions for a weight parameter of each of at least one layer of the neural network, determining a dimension value and a corresponding target value in each dimension of the weight parameter, and padding the weight parameter in a case where the dimension value in at least one dimension of the weight parameter is less than the corresponding target value, the dimension value in each dimension of the weight parameter after the padding being equal to the corresponding target value.

Abstract: Disclosed is an apparatus for performing a convolution operation in a convolutional neural network. The apparatus may comprise a selector for selecting one or more nonzero elements of a weight parameter, a selector for selecting a data item(s) corresponding to selected nonzero elements in input feature data, and a calculator unit for performing an operation. The apparatus may realize the convolution operation in a sparsified convolutional neural network efficiently through the hardware.

Abstract: Disclosed are a circuit unit, a circuit module and an apparatus for data statistics. The circuit unit comprises a first register and a second register, and stores data received via a first input terminal in the first register in a case where a first control terminal receives a valid control signal, stores data received via a second input terminal in the second register in a case where a second control terminal receives a valid control signal, and increases the value of data stored in the second register by 1 in a case where a third control terminal receives a valid control signal. The circuit module comprises one or more such circuit units, and the apparatus comprises one or more such circuit modules. The circuit module or the apparatus may use smaller resource and smaller power consumption to complete data statistics.

Abstract: Disclosed is a method and an apparatus for recognizing speech, and the method comprises: separating an input audio signal into at least two separated signals; generating a denoised signal at a current frame; performing a preliminary recognition on each interesting signal at the current frame; and performing a recognition decision according to a recognition score of each interesting signal at the current frame. The method and apparatus of the present disclosure deeply integrate an array signal processing and a speech recognition and use multiway recognitions such that a good recognition rate may be obtained even in a case of a low signal-to-noise ratio.

Abstract: A method and an apparatus for reducing noise of mixed signal are disclosed. The method includes: separating a collected mixed signal to obtain a first signal and a second signal; selecting one of the first signal and the second signal as a current reference signal, and the other as a current expected signal; and performing adaptive filtering based on the selected current reference signal and the selected current expected signal. By the method and the apparatus, the noise can be reduced significantly or removed in a case where reference signal cannot be directly obtained from a hardware.

Abstract: Disclosed are a method and an apparatus for blind signal separation and an electronic device. The method includes modeling a sound source with a complex Gaussian distribution to determine a probability density distribution of the sound source; updating a blind signal separation model based on the probability density distribution; and separating an audio signal with the updated blind signal separation model to obtain a plurality of separated output signals. In this way, the blind signal separation model may be updated through the probability density distribution of the sound source obtained based on the complex Gaussian distribution, thereby effectively improving separation performance of a blind signal separation algorithm in specific scenario.

Abstract: A sound-processing apparatus comprises at least one pair of sound transducers. A first sound transducer may receive an audio source signal and output a first sound signal according to the audio source signal. A second sound transducer may receive the audio source signal and output a second sound signal according to the audio source signal, the second sound signal having opposite phase from the first sound signal. A difference between the second and first sound signal amplitudes may be less than or equal to an amplitude threshold value. A sound acquisition device may acquire a sound signal. Path-characteristic differences between amplitude-frequency characteristics of a first sound path from the first sound transducer to the sound acquisition device and a second sound path from the second sound transducer to the sound acquisition device may be less than or equal to a first characteristic threshold value.