Abstract: The present disclosure relates to the field of audio enhancement, and in particular to methods, devices and software for supervised training of a machine learning model, MLM, the MLM trained to enhance a degraded audio signal by calculating gains to be applied to frequency bands of the degraded audio signal. The present disclosure further relates to methods, devices and software for use of such a trained MLM.

Abstract: A method for dereverberating audio signals is provided. In some implementations, the method involves obtaining a real acoustic impulse response (AIR); identifying a first portion of the real AIR corresponding to early reflections of a direct sound and a second portion of the real AIR that corresponding to late reflections of the direct sound; generating one or more synthesized AIRs by modifying the first portion of the real AIR and/or the second portion of the real AIR; and using the real AIR and the one or more synthesized AIRs to generate a plurality of training samples, each training sample comprising an input audio signal and a reverberated audio signal, wherein the reverberated audio signal is generated based on the input audio signal and one of the real AIR or one of the one or more synthesized AIRs, which plurality of training samples are used to train a machine learning model.

Abstract: A system for mitigating over-suppression of speech and other non-noise signals is disclosed. In some embodiments, a system is programmed to train a first machine learning model for speech detection or enhancement using a non-linear, asymmetric loss function that penalizes speech over-suppression more than speech under-suppression. The first machine learning model is configured to receive an audio signal and generate a mask indicating an amount of speech present in the audio signal. The mask can be adjusted to remedy sharp voice decay resulting from speech over-suppression. The system is also programmed to train a second machine learning model for laughter or applause detection. The system is further programmed to improve the quality of a new audio signal by applying an adjusted mask to the new audio signal except for the portions of the audio signal that have been identified as corresponding to laughter or applause.

Abstract: Described herein is a method for Convolutional Neural Network (CNN) based speech source separation, wherein the method includes the steps of: (a) providing multiple frames of a time-frequency transform of an original noisy speech signal; (b) inputting the time-frequency transform of said multiple frames into an aggregated multi-scale CNN having a plurality of parallel convolution paths; (c) extracting and outputting, by each parallel convolution path, features from the input time-frequency transform of said multiple frames; (d) obtaining an aggregated output of the outputs of the parallel convolution paths; and (e) generating an output mask for extracting speech from the original noisy speech signal based on the aggregated output. Described herein are further an apparatus for CNN based speech source separation as well as a respective computer program product comprising a computer-readable storage medium with instructions adapted to carry out said method when executed by a device having processing capability.

Abstract: Disclosed is a vacuum coating device, including a heating apparatus, a loading plate apparatus, a spraying apparatus, a processing chamber and a plurality of valve chambers. The heating apparatus is configured to heat the processing chamber to a temperature that reaches a reaction temperature of the object to be coated and the processing gas, the loading plate apparatus is configured to load the object to be coated, the processing chamber includes a first coating vacuum chamber, a second coating vacuum chamber and a loading plate conversion vacuum chamber, and the loading plate conversion vacuum chamber is configured to convert a coating surface of an object to be coated to achieve a function of coating both sides of the object to be coated on a closed production line, and to improve production quality of the object to be coated.

Abstract: A method for enhancing audio signals is provided. In some implementations, the method involves (a) obtaining a training set comprising a plurality of training samples, each training sample comprising a distorted audio signal and a clean audio signal. In some implementations, the method involves (b), for a training sample of the plurality of training samples: obtaining a frequency-domain representation of the distorted audio signal; providing the frequency-domain representation to a convolutional neural network (CNN) comprising a plurality of convolutional layers and to a recurrent element, wherein an output of the recurrent element is provided to a subset of the plurality of convolutional layers; generating a predicted enhancement mask, wherein the CNN generates the predicted enhancement mask; generating a predicted enhanced audio signal based on the predicted enhancement mask; and updating weights associated with the CNN and the recurrent element based on the predicted enhanced audio signal.

Abstract: The present disclosure relates to the field of audio enhancement, and in particular to methods, devices and software for supervised training of a machine learning model, MLM, the MLM trained to enhance a degraded audio signal by calculating gains to be applied to frequency bands of the degraded audio signal. The present disclosure further relates to methods, devices and software for use of such a trained MLM.

Abstract: Described herein is a method of low-bitrate coding of audio data and generating enhancement metadata for controlling audio enhancement of the low-bitrate coded audio data at a decoder side, including the steps of: (a) core encoding original audio data at a low bitrate to obtain encoded audio data; (b) generating enhancement metadata to be used for controlling a type and/or amount of audio enhancement at the decoder side after core decoding the encoded audio data; and (c) outputting the encoded audio data and the enhancement metadata. Described is further an encoder configured to perform said method. Described is moreover a method for generating enhanced audio data from low-bitrate coded audio data based on enhancement metadata and a decoder configured to perform said method.

Abstract: The present disclosure relates to the field of audio enhancement, and in particular to methods, devices and software for supervised training of a machine learning model, MLM, the MLM trained to enhance a degraded audio signal by calculating gains to be applied to frequency bands of the degraded audio signal. The present disclosure further relates to methods, devices and software for use of such a trained MLM.

Abstract: Described herein is a method for Convolutional Neural Network (CNN) based speech source separation, wherein the method includes the steps of: (a) providing multiple frames of a time-frequency transform of an original noisy speech signal; (b) inputting the time-frequency transform of said multiple frames into an aggregated multi-scale CNN having a plurality of parallel convolution paths; (c) extracting and outputting, by each parallel convolution path, features from the input time-frequency transform of said multiple frames; (d) obtaining an aggregated output of the outputs of the parallel convolution paths; and (e) generating an output mask for extracting speech from the original noisy speech signal based on the aggregated output. Described herein are further an apparatus for CNN based speech source separation as well as a respective computer program product comprising a computer-readable storage medium with instructions adapted to carry out said method when executed by a device having processing capability.

Abstract: A method of manufacturing a semiconductor device is provided. The method includes forming a channel layer and an active layer over a substrate; forming a doped epitaxial layer over the active layer; patterning the doped epitaxial layer, the active layer, and the channel layer to form a fin structure comprising a doped epitaxial fin portion, an active fin portion below the doped epitaxial fin portion, and a channel fin portion below the active fin portion; removing the doped epitaxial fin portion; and forming a gate electrode at least partially extending along a sidewall of the fin structure to form a Schottky barrier between the gate electrode and the fin structure after removing the doped epitaxial fin portion.

Abstract: Described herein is a method of low-bitrate coding of audio data and generating enhancement metadata for controlling audio enhancement of the low-bitrate coded audio data at a decoder side, including the steps of: (a) core encoding original audio data at a low bitrate to obtain encoded audio data; (b) generating enhancement metadata to be used for controlling a type and/or amount of audio enhancement at the decoder side after core decoding the encoded audio data; and (c) outputting the encoded audio data and the enhancement metadata. Described is further an encoder configured to perform said method. Described is moreover a method for generating enhanced audio data from low-bitrate coded audio data based on enhancement metadata and a decoder configured to perform said method.

Abstract: A method of manufacturing a semiconductor device is provided. The method includes forming a channel layer and an active layer over a substrate; forming a doped epitaxial layer over the active layer; patterning the doped epitaxial layer, the active layer, and the channel layer to form a fin structure comprising a doped epitaxial fin portion, an active fin portion below the doped epitaxial fin portion, and a channel fin portion below the active fin portion; removing the doped epitaxial fin portion; and forming a gate electrode at least partially extending along a sidewall of the fin structure to form a Schottky barrier between the gate electrode and the fin structure after removing the doped epitaxial fin portion.

Abstract: A semiconductor device includes a substrate, a channel layer, an active layer, and a gate electrode. The channel layer has a fin portion over the substrate. The active layer is over at least the fin portion of the channel layer. The active layer is configured to cause a two-dimensional electron gas (2DEG) to be formed in the channel layer along an interface between the channel layer and the active layer. The gate electrode is in contact with a sidewall of the fin portion of the channel layer.

Abstract: The present application provides a method and a system for identifying a user device. The method comprises acquiring multiple device records ranked in chronological order; calculating time-domain feature information of a changed attribute value in the device records; calculating a difference value between values of an attribute before and after change according to the time-domain feature information of the changed attribute value in the device records; calculating a similarity of the device records according to difference values of all attributes in the device records; and identifying whether the device records belong to an identical device according to the similarity of the device records. As compared with current systems, the identification of a user device is done based on all the attributes of a device in which a similarity of device records is precisely calculated through a similarity weight that is constantly adjusted time sequences, thereby effectively reducing the misjudgment rate.

Abstract: A semiconductor device includes a substrate, a channel layer, an active layer, and a gate electrode. The channel layer has a fin portion over the substrate. The active layer is over at least the fin portion of the channel layer. The active layer is configured to cause a two-dimensional electron gas (2DEG) to be formed in the channel layer along an interface between the channel layer and the active layer. The gate electrode is in contact with a sidewall of the fin portion of the channel layer.

Abstract: A method for making a cathode active material of a lithium ion battery is disclosed. In the method, LiMPO4 particles and LiNPO4 particles are provided. The LiMPO4 particles and LiNPO4 particles both are olivine type crystals belonged to a pnma space group of an orthorhombic crystal system, wherein M represents Fe, Mn, Co, or Ni, N represents a metal element having a +2 valence, and N is different from M. The LiMPO4 particles and the LiNPO4 particles are mixed together to form a precursor. The precursor is calcined to form LiMxN1-xPO4 particles, wherein 0<x<1.

Abstract: The present application provides a method and a system for identifying a user device. The method comprises acquiring multiple device records ranked in chronological order; calculating time-domain feature information of a changed attribute value in the device records; calculating a difference value between values of an attribute before and after change according to the time-domain feature information of the changed attribute value in the device records; calculating a similarity of the device records according to difference values of all attributes in the device records; and identifying whether the device records belong to an identical device according to the similarity of the device records. As compared with current systems, the identification of a user device is done based on all the attributes of a device in which a similarity of device records is precisely calculated through a similarity weight that is constantly adjusted time sequences, thereby effectively reducing the misjudgment rate.

Abstract: A method for making a lithium ion battery anode active material comprising: providing silicon particles and a silane coupling agent, wherein the silane coupling agent comprises a hydrolysable functional group and an organic functional group; mixing the silicon particles and the silane coupling agent in water to obtain a first mixture; adding a monomer or oligomer to the first mixture to obtain a second mixture, the surfaces of the silicon particles being coated with a polymer layer by in situ polymerization method to obtain silicon polymer composite material, the monomer or the oligomer reacting with the organic functional group of the silane coupling agent in a polymerization, thereby a generated polymer layer being chemically grafted on the surfaces of the silicon particles; and heating the silicon polymer composite material to carbonize the polymer layer to form a carbon layer coated on the surfaces of the silicon particles.

Abstract: A method for making a cathode active material of a lithium ion battery is disclosed. In the method, LiMPO4 particles and LiNPO4 particles are provided. The LiMPO4 particles and LiNPO4particles both are olivine type crystals belonged to a pnma space group of an orthorhombic crystal system, wherein M represents Fe, Mn, Co, or Ni, N represents a metal element having a +2 valence, and N is different from M. The LiMPO4 particles and the LiNPO4 particles are mixed together to form a precursor. The precursor is calcined to form LiMxN1-xPO4 particles, wherein 0<x<1.