-
公开(公告)号:EP4425382A1
公开(公告)日:2024-09-04
申请号:EP22892105.2
申请日:2022-11-11
发明人: YE, Deshi , CHEN, Songyang , XU, Chen , LI, Rong
IPC分类号: G06N3/08
CPC分类号: G06F18/25 , G06N5/04 , G06F18/214 , G06N3/08 , G06N3/06 , G06N3/084 , G06N3/0464
摘要: This application provides a model training method and a communication apparatus. The method includes: An ith device in a kth group of devices performs n*M times of model training, where the ith device completes model parameter exchange with at least one other device in the kth group of devices in every M times of model training, M is a quantity of devices in the kth group of devices, M is greater than or equal to 2, and n is an integer; and the ith device sends a model Mi,n*M to a target device, where the model Mi,n*M is a model obtained by the ith device by completing the n*M times of model training. According to the solutions provided in this application, a convergence speed and utilization of data and computing power in a model training process can be improved, and requirements on computing, storage, and communication capabilities of a device can be reduced.
-
公开(公告)号:EP4318310A1
公开(公告)日:2024-02-07
申请号:EP22787608.3
申请日:2022-04-14
发明人: ZHANG, Gongzheng , XU, Chen , LI, Rong , WANG, Jian , WANG, Jun
摘要: This application provides a distributed learning method and apparatus, to combine wireless communication with distributed learning, which can save resources, and improve performance of distributed learning in a wireless environment. The method includes: A first node processes first data by using a first data model to obtain first intermediate data, and sends the first intermediate data to a second node through a first channel. The first channel is updated based on error information of second intermediate data, information about the first channel, and the first intermediate data. The second intermediate data is a result of transmitting the first intermediate data to the second node through the first channel. The first channel is a channel between the first node and the second node.
-
公开(公告)号:EP4210252A1
公开(公告)日:2023-07-12
申请号:EP21874417.5
申请日:2021-09-27
发明人: XU, Chen , ZHANG, Gongzheng , WANG, Jian , LI, Rong , WANG, Jun , TONG, Wen
摘要: Embodiments of this application disclose a neural network-based communication method and a related apparatus. Specifically, joint training optimization is performed on an encoding neural network used by a transmit end and a decoding neural network used by a receive end. A first neural network in the encoding neural network reuses the decoding neural network and a parameter of the decoding neural network. In this way, the decoding neural network provides a part of prior information for the encoding neural network, so that the encoding neural network can more accurately predict a benefit of an action, thereby reducing a variance of gradient estimation, accelerating a network convergence speed, and improving convergence performance.
-
公开(公告)号:EP4054148A1
公开(公告)日:2022-09-07
申请号:EP20894398.5
申请日:2020-11-02
发明人: WANG, Jian , XU, Chen , HUANGFU, Yourui , LI, Rong , WANG, Jun
IPC分类号: H04L29/08
摘要: This application provides a scheduling method and apparatus, to optimize a MAC scheduling algorithm. The method includes: a network device uses a scheduling model applicable to K users for one or more times, and determines a scheduled terminal device based on scheduling weights output by the scheduling model for one or more times. Each time the scheduling model is used, status information of K terminal devices in n to-be-scheduled terminal devices is input to the scheduling model, and the scheduling model outputs scheduling weights respectively corresponding to the K terminal devices, where K is an integer greater than 1, and n is an integer greater than K. Therefore, regardless of a quantity of to-be-scheduled terminal devices in a communication system, the network device may reuse the scheduling model without re-establishing a scheduling model, so that the scheduling model is expandable and applicable to scenarios in which quantities of to-be-scheduled terminal devices are different.
-
公开(公告)号:EP3910830A1
公开(公告)日:2021-11-17
申请号:EP21163022.3
申请日:2018-05-04
发明人: HUANG, Lingchen , DAI, Shengchen , XU, Chen , QIAO, Yunfei , LI, Rong
摘要: This application provides a decoding method and apparatus. The method includes performing polar decoding based on a CRC polynomial to obtain a polar-decoded bit sequence, wherein the polar-decoded bit sequence comprises 11 CRC bits. Based on an improved CRC polynomial, a decoding satisfying an FAR requirement is implemented.
-
公开(公告)号:EP3457607A1
公开(公告)日:2019-03-20
申请号:EP18771407.6
申请日:2018-03-22
发明人: XU, Chen , ZHANG, Gongzheng , LI, Rong , ZHANG, Huazi , HUANG, Lingchen
IPC分类号: H04L1/00
摘要: Embodiments of the application provide a method and device for polar code rate matching. The method includes: polar encoding according to a mother code length N and a rate matching manner, to obtain encoded bits after the polar encoding, where the rate matching manner is a first rate matching manner or a second rate matching manner, and N is a positive integer; storing, into a cyclic cache according to a bit storage order corresponding to the rate matching manner, the encoded bits, wherein the first rate matching manner and the second rate matching manner are corresponding to a same bit storage order; and reading, from the cyclic cache according to a bit reading order corresponding to the rate matching manner, an output sequence obtained after rate matching, wherein the first rate matching manner and the second rate matching manner are corresponding to different bit reading orders. According to the embodiments, hardware complexity and an area occupied by hardware can be reduced.
-
公开(公告)号:EP3422620A1
公开(公告)日:2019-01-02
申请号:EP18746067.0
申请日:2018-03-23
发明人: HUANG, Lingchen , ZHANG, Gongzheng , XU, Chen , ZHANG, Chaolong , WANG, Jun
IPC分类号: H04L1/00
摘要: Embodiments of this application provide a method and an apparatus for constructing a sequence of polar codes. The method includes: reading a tectonic sequence P' whose length is the same as an encoding code length of a to-be-constructed sequence of polar codes from a tectonic sequence P , where the tectonic sequence P is generated based on a reliability ranking sequence Q and a rate matching rule; demapping the tectonic sequence P ' to a reliability ranking sequence Q ' based on rate matching; and constructing the sequence of polar codes based on the reliability ranking sequence Q '. According to this application that is implemented, a rate matching process and a process of constructing a sequence of polar codes are combined, thereby effectively improving efficiency in constructing a sequence of polar codes, and better adapting to the rate matching process.
-
公开(公告)号:EP4266633A1
公开(公告)日:2023-10-25
申请号:EP21909442.2
申请日:2021-12-22
发明人: ZHANG, Gongzheng , HU, Bin , XU, Chen , LI, Rong , MA, Jianglei
摘要: Embodiments of this application provide a communication method and an apparatus, which are applied to the field of communications technologies. The communication method includes: A second device receives policy related information from M first devices; the second device obtains transmission decisions of the M first devices based on the policy related information by using a second neural network; the second device updates the second neural network based on reward information, and sends, to the M first devices, information for updating a first neural network, and the third device obtains second update parameter information of the first neural network based on the first update parameter information of the first neural network of the M first devices, and sends the second update parameter information of the first neural network to the M first devices, so that the first device may update the first neural network based on the second update parameter information. In the method in embodiments of this application, the second update parameter information of the first neural network is obtained in a training process between the third device or the second device and the M first devices, so that training overheads can be reduced.
-
公开(公告)号:EP4340283A1
公开(公告)日:2024-03-20
申请号:EP22810466.7
申请日:2022-05-20
发明人: WANG, Jian , XU, Chen , ZHANG, Gongzheng , QIAO, Yunfei , LI, Rong , WANG, Jun
摘要: Embodiments of this application disclose a data transmission method and a related apparatus, to train, at low signaling overheads, machine learning models deployed at a transmitting side and a receiving side in an end-to-end communication system. The method in embodiments of this application is applied to a communication system including a first communication apparatus and a second communication apparatus, a first machine learning model is deployed in the first communication apparatus, and a second machine learning model is deployed in the second communication apparatus. The method includes: obtaining first information that carries indication information of both a first transmission resource and a second transmission resource, where the first transmission resource is for the first communication apparatus to transmit a first output of the first machine learning model to the second communication apparatus; and the second transmission resource is for the first communication apparatus to receive first feedback data that is from the second communication apparatus and that includes a first gradient, and the first gradient is for updating the first machine learning model. The first communication apparatus transmits the first output to the second communication apparatus on the first transmission resource.
-
公开(公告)号:EP4322065A1
公开(公告)日:2024-02-14
申请号:EP22787460.9
申请日:2022-04-08
发明人: XU, Chen , ZHANG, Gongzheng , WANG, Jian , LI, Rong , WANG, Jun
摘要: This application provides a gradient transmission method and a related apparatus. In this method, a first communication apparatus receives training data, and determines a first intermediate gradient based on the training data. The first intermediate gradient is used to update a parameter of a first neural network located in a second communication apparatus. The first communication apparatus maps the first intermediate gradient to an air interface resource to generate a first gradient signal, and sends the first gradient signal to the second communication apparatus. The first gradient signal includes one or more first gradient symbols, and each of the first gradient symbols is corresponding to one or more gradient values. The first communication apparatus maps the first intermediate gradient to the air interface resource for transmission, that is, implements reverse transmission of the first intermediate gradient by sending the first gradient signal to the second communication apparatus, to facilitate implementing joint training of the first communication apparatus and the second communication apparatus.
-
-
-
-
-
-
-
-
-