Tencent Technology (shenzhen) Company Limited

Huawei Device Co., Ltd.

Huawei Technologies Co., Limited

Byd Company Limited

Shenzhen Goodix Technology Co., Ltd.

Huawei Device (shenzhen) Co., Ltd.

Sanechips Technology Co., Ltd.

Tencent Technology Company Limited

Na Li

Hui Jin

Qiang Yi

Jun Zhang

Guodong Xue

Jun Wang

Tao Li

Jie Shi

Qi Yu

Guoliang Li

Xinshan Wang

Bengang Yi

Deyi Tuo

Hongjing Guo

Yuan Tao

Can Wang

Hu Zhu

Yubin Wang

Junda Ma

Yingzhong Chen

Xiaobo Mao

Bo Song

Shining Jiang

A voice identity feature extractor training method includes extracting a voice feature vector of training voice. The method may include determining a corresponding I-vector according to the voice feature vector of the training voice. The method may include adjusting a weight of a neural network model by using the I-vector as a first target output of the neural network model, to obtain a first neural network model. The method may include obtaining a voice feature vector of target detecting voice and determining an output result of the first neural network model for the voice feature vector of the target detecting voice. The method may include determining an I-vector latent variable. The method may include estimating a posterior mean of the I-vector latent variable, and adjusting a weight of the first neural network model using the posterior mean as a second target output, to obtain a voice identity feature extractor.

Voice identity feature extractor and classifier training

Embodiments of this application disclose a voiceprint recognition method performed by a computer. After obtaining a to-be-recognized target voice message, the computer obtains target feature information of the target voice message by using a voice recognition model, the voice recognition model being obtained through training according to a first loss function and a second loss function. Next, the computer determines a voiceprint recognition result according to the target feature information and registration feature information, the registration feature information being obtained from a voice message of a to-be-recognized object using the voiceprint recognition model. The normalized exponential function and the centralization function are used for jointly optimizing the voice recognition model, and can reduce an intra-class variation between depth features from the same speaker. The two functions are used for simultaneously supervising and learning the voice recognition model, and enable the depth feature to have better discrimination, thereby improving recognition performance.

Voiceprint recognition method, model training method, and server

Some embodiments of the present disclosure provide a method and an apparatus for forming a differential beam, a method and an apparatus for processing a signal, and a chip. The method for forming a differential beam includes: obtaining a differential beam forming signal according to an input signal acquired by two microphones in a microphone array (); and performing a nonlinear adjustment on at least an amplitude of the differential beam forming signal based on a distance between the two microphones and a signal frequency of the input signal to obtain the adjusted differential beam forming signal (). With the above solution, a constant beam characteristic of the differential beam forming signal can be ensured as much as possible for microphone arrays of different specifications.

Method and apparatus for forming differential beam, method and apparatus for processing signal, and chip

A method and device for transmitting data are provided. The method includes: determining, after receiving a first message sent by a user equipment, whether first indication information is set, where the first indication information indicates that a server buffers downlink data to be sent to the user equipment; and when determining that the first indication information is set, retaining a signaling connection to the user equipment, and triggering the server that buffers the downlink data to send the downlink data to the user equipment. The method and device for transmitting data according to the embodiments of the present invention can save power of a user equipment.

Method and device for transmitting data

The present invention provides M2M data querying and invoking methods, querying and invoking devices, and a system. The querying method includes: sending a data invoking request to an infrastructure node common services entity infrastructure node CSE, where the data invoking request includes an identifier AE/CSE ID of an application entity/a common services entity that provides to-be-invoked data and a data demand of the to-be-invoked data, where the data demand includes a data type demand, a time demand, or a combination thereof; and receiving data meeting the data demand and returned by the infrastructure node CSE.

M2M data querying and invoking methods, querying and invoking devices, and system

Embodiments of the present invention provide a method and device for transmitting data. The method includes: determining, after receiving a first message sent by a user equipment, whether first indication information is set, where the first indication information indicates that a server buffers downlink data to be sent to the user equipment; and when determining that the first indication information is set, retaining a signaling connection to the user equipment, and triggering the server that buffers the downlink data to send the downlink data to the user equipment. The method and device for transmitting data according to the embodiments of the present invention can save power of a user equipment.

In transmitting a device trigger message operation, a core network device receives a device trigger message sent by a first device, where the device trigger message includes information that indicates validity time of the device trigger message; when it is determined that a timer exists, compares the validity time with duration of the timer; if the validity time is greater than the duration of the timer, sends the device trigger message to a user equipment UE after the timer times out; and if the validity time is not greater than the duration of the timer, sends a message to the first device and/or discarding the device trigger message, where the message is used to indicate that triggering is rejected, or transmission fails, or device triggering fails.

Method and core network device for transmitting device trigger message

Car undercarriage

Growing community of inventors

Shenzhen, China

Na Li