专利快速检索-快速检索全球专利，免费商用专利数据库-IPRDB

1. 发明申请

WO2019155064A1 DATA COMPRESSION USING JOINTLY TRAINED ENCODER, DECODER, AND PRIOR NEURAL NETWORKS 审中-公开
公开(公告)号：WO2019155064A1
公开(公告)日：2019-08-15
申请号：PCT/EP2019/053322
申请日：2019-02-11
申请人： DEEPMIND TECHNOLOGIES LIMITED
发明人： MENICK, Jacob Lee , GRAVES, Alexander Benjamin
IPC分类号： G06N3/04 , G06N3/08 , G06N3/00 , G10L19/00 , H03M7/40
CPC分类号： G06N3/0454 , G06N3/006 , G06N3/084 , G06N3/088 , G10L19/00 , H03M7/40
摘要： Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for training an encoder neural network, a decoder neural network, and a prior neural network, and using the trained networks for generative modeling, data compression, and data decompression. In one aspect, a method comprises: providing a given observation as input to the encoder neural network to generate parameters of an encoding probability distribution; determining an updated code for the given observation; selecting a code that is assigned to an additional observation; providing the code assigned to the additional observation as input to the prior neural network to generate parameters of a prior probability distribution; sampling latent variables from the encoding probability distribution; providing the latent variables as input to the decoder neural network to generate parameters of an observation probability distribution; and determining gradients of a loss function.

2. 发明申请

WO2019145082A1 A METHOD FOR COLLABORATIVE MACHINE LEARNING OF ANALYTICAL MODELS 审中-公开
公开(公告)号：WO2019145082A1
公开(公告)日：2019-08-01
申请号：PCT/EP2018/084201
申请日：2018-12-10
申请人： SIEMENS AKTIENGESELLSCHAFT
发明人： SOLER GARRIDO, Josep , KROMPASS, Denis , FISCHER, Jan-Gregor
IPC分类号： G06Q10/06 , G06Q50/04 , G06Q50/06 , G06N3/04
CPC分类号： G06Q10/06 , G06N3/0454 , G06Q50/04 , G06Q50/06
摘要： A method for machine learning of analytical models, AMs, comprising core model components, CMCs, shared between tasks, t, of different customers (A, B) and comprising specialized model components, SMCs, specific to customer tasks, t, of individual customers (A, B), wherein the machine learning of the analytical models, AMs, is performed collaboratively based on local data, LD, provided by machines (3A, 3B) of the customer premises (2A, 2B) of different customers (A, B) without the local data, LD, leaving the respective customer premises (2A, 2B).

3. 发明申请

WO2019126755A1 GENERATING AND CLASSIFYING TRAINING DATA FOR MACHINE LEARNING FUNCTIONS 审中-公开
公开(公告)号：WO2019126755A1
公开(公告)日：2019-06-27
申请号：PCT/US2018/067298
申请日：2018-12-21
申请人： FUGRO N.V. , MARTIN, Dickie Andrew , GAUDET, Chase John
发明人： MARTIN, Dickie Andrew , GAUDET, Chase John
IPC分类号： G06N20/00
CPC分类号： G06N3/0445 , G05D1/00 , G06N3/0454 , G06N3/084 , G06N3/086
摘要： Disclosed are systems, methods and computer readable mediums for generating and applying training data to training data computer learning systems and for autonomous operation of surface and sub-surface vessels. The systems, methods and computer readable mediums can include a data source, the data source providing one or more of vehicle data, operational data, and environmental data and can further include a processor, the processor analyzing at least data from the data source and generating one or more control outputs based at least in part on the analysis.

4. 发明申请

WO2019125708A1 REDUCING POWER CONSUMPTION IN A NEURAL NETWORK ENVIRONMENT USING DATA MANAGEMENT 审中-公开
公开(公告)号：WO2019125708A1
公开(公告)日：2019-06-27
申请号：PCT/US2018/062690
申请日：2018-11-28
申请人： MICROSOFT TECHNOLOGY LICENSING, LLC
发明人： AMBARDEKAR, Amol Ashok , MCBRIDE, Chad Balling , PETRE, George , CEDOLA, Kent D. , WALL, Larry Marvin
IPC分类号： G06N3/063 , G06N3/04
CPC分类号： G06F1/3243 , G06F1/3275 , G06N3/04 , G06N3/0454 , G06N3/063
摘要： Techniques to provide for improved (i.e., reduced) power consumption in an exemplary neural network (NN) and/or Deep Neural Network (DNN) environment using data management. Improved power consumption in the NN/DNN may be achieved by reducing a number of bit flips needed to process operands associated with one or more storages. Reducing the number bit flips associated with the NN/DNN may be achieved by multiplying an operand associated with a first storage with a plurality of individual operands associated with a plurality of kernels of the NN/DNN. The operand associated with the first storage may be neuron input data and the plurality of individual operands associated with the second storage may be weight values for multiplication with the neuron input data. The plurality of kernels may be arranged or sorted and subsequently processed in a manner that improves power consumption in the NN/DNN.

5. 发明申请

WO2019125453A1 TRAINING A CONVOLUTIONAL NEURAL NETWORK USING TASKIRRELEVANT DATA 审中-公开
公开(公告)号：WO2019125453A1
公开(公告)日：2019-06-27
申请号：PCT/US2017/067766
申请日：2017-12-21
申请人： SIEMENS AKTIENGESELLSCHAFT
发明人： MANJUNATHA, Varun , GEORGAKIS, Georgios , PENG, Kuan-Chuan , WU, Ziyan , ERNST, Jan
IPC分类号： G06K9/62
CPC分类号： G06K9/6273 , G06K9/6288 , G06N3/0454 , G06N3/08
摘要： Examples of techniques for training a convolutional neural network are disclosed. In one example implementation according to aspects of the present disclosure, a computer-implemented method includes receiving, by a processing device, a taskirrelevant image pair comprising an irrelevant depth image and an irrelevant RGB image. The method further includes feeding, by the processing device, the irrelevant depth image into the source CNN. The method further includes feeding, by the processing device, the irrelevant RGB image into a target CNN. The method further includes performing, by the processing device, a first Euclidean loss to encourage features of the irrelevant depth image and features of the irrelevant RGB image to be similar. The method further includes training, by the processing device, the source CNN based at least in part on the first Euclidean loss between the features of the irrelevant depth image and the features of the irrelevant RGB image.

6. 发明申请

WO2019115236A1 INDEPENDENT AND DEPENDENT READING USING RECURRENT NETWORKS FOR NATURAL LANGUAGE INFERENCE 审中-公开
公开(公告)号：WO2019115236A1
公开(公告)日：2019-06-20
申请号：PCT/EP2018/082915
申请日：2018-11-29
申请人： KONINKLIJKE PHILIPS N.V.
发明人： GHAEINI, Reza , AL HASAN, Sheikh Sadid , FARRI, Oladimeji Feyisetan
IPC分类号： G06N3/04
CPC分类号： G06N3/0445 , G06N3/0454
摘要： Techniques disclosed herein related to independent and dependent reading using recurrent networks for natural language inference. In various embodiments, data indicative of a premise (310) and data indicative of a hypothesis (312) form a natural language inference classification pair. For example, the data indicative of a premise can be processed independently using a third recurrent network (318) and data indicative of a hypothesis can be processed independently using a first recurrent network (314). Similarly, data indicative of a premise can be processed dependently using a second recurrent network (316) including data indicative of a hypothesis processed independently. Additionally, data indicative of a hypothesis can be processed dependently using a fourth recurrent network (320) including data indicative of a premise processed independently. Independent and dependent premise data can be pooled (334) together. Independent and dependent hypothesis data can be pooled (336) together.

7. 发明申请

WO2019113354A1 SYSTEMS FOR CONSTRUCTING HIERARCHICAL TRAINING DATA SETS FOR USE WITH MACHINE-LEARNING AND RELATED METHODS THEREFOR 审中-公开
公开(公告)号：WO2019113354A1
公开(公告)日：2019-06-13
申请号：PCT/US2018/064308
申请日：2018-12-06
申请人： ZERO MASS WATER, INC.
发明人： FRIESEN, Cody , JOHNSON, Paul , LORZEL, Heath , SALLOUM, Kamil , GOLDBERG, Jonathan , FRIESEN, Grant , HORWITZ, Jason
IPC分类号： G05B13/02 , G05B17/02 , G06N20/00
CPC分类号： G06N20/00 , G05B13/0265 , G05B17/02 , G06N3/0454 , G06N3/08
摘要： Some embodiments include a system operable to construct hierarchical training data sets for use with machine-learning for multiple controlled devices. Other embodiments of related systems and methods are also provided.

8. 发明申请

WO2019108923A1 NEURAL ARCHITECTURE SEARCH USING A PERFORMANCE PREDICTION NEURAL NETWORK 审中-公开
公开(公告)号：WO2019108923A1
公开(公告)日：2019-06-06
申请号：PCT/US2018/063293
申请日：2018-11-30
申请人： GOOGLE LLC
发明人： HUA, Wei , ZOPH, Barret , SHLENS, Jonathon , LIU, Chenxi , HUANG, Jonathan , LI, Jia , LI, Fei-Fei , MURPHY, Kevin Patrick
IPC分类号： G06N3/08 , G06N3/04
CPC分类号： G06N3/082 , G06N3/0445 , G06N3/0454
摘要： A method for determining an architecture for a task neural network configured to perform a particular machine learning task is described. The method includes obtaining data specifying a current set of candidate architectures for the task neural network; for each candidate architecture in the current set: processing the data specifying the candidate architecture using a performance prediction neural network having multiple performance prediction parameters, the performance prediction neural network being configured to process the data specifying the candidate architecture in accordance with current values of the performance prediction parameters to generate a performance prediction that characterizes how well a neural network having the candidate architecture would perform after being trained on the particular machine learning task; and generating an updated set of candidate architectures by selecting one or more of the candidate architectures in the current set based on the performance predictions for the candidate architectures in the current set.

9. 发明申请

WO2019104695A1 用于神经网络的运算装置、芯片、设备及相关方法审中-公开
公开(公告)号：WO2019104695A1
公开(公告)日：2019-06-06
申请号：PCT/CN2017/114079
申请日：2017-11-30
申请人：深圳市大疆创新科技有限公司
发明人：韩峰 , 李鹏 , 谷骞
IPC分类号： G06F9/28
CPC分类号： G06F7/52 , G06N3/0454
摘要：提供一种用于神经网络的运算装置、芯片与设备，所述运算装置包括：控制单元与乘累加单元组，该乘累加单元组包括滤波器寄存器与多个计算单元，该滤波器寄存器与该多个计算单元连接；该控制单元用于，生成控制信息，并向该计算单元发送该控制信息；该滤波器寄存器用于，缓存待进行乘累加运算的滤波器权重值；该计算单元用于，缓存待进行乘累加运算的输入特征值，根据接收的该控制信息对该滤波器权重值与该输入特征值进行乘累加运算。通过一个控制单元控制所有的计算单元，可以降低控制单元的设计复杂度；通过多个计算单元共用一个滤波器寄存器，可以减少所需的缓存大小。

10. 发明申请

WO2019096521A1 A METHOD FOR IMPROVING USER AUTHENTICATION PERFORMED BY A COMMUNICATION DEVICE 审中-公开
公开(公告)号：WO2019096521A1
公开(公告)日：2019-05-23
申请号：PCT/EP2018/078300
申请日：2018-10-16
申请人： GEMALTO SA
发明人： FAURE, Frédéric
IPC分类号： G06F21/31 , G06F21/32 , H04L29/06 , H04W12/06 , G06N3/08
CPC分类号： G06F21/31 , G06F21/32 , G06N3/0454 , G06N3/088 , H04L63/0861 , H04L2463/082 , H04W12/00508 , H04W12/06
摘要： This invention relates to a method for improving user authentication efficiency performed by a communication device belonging to an authentication system, the communication device (200) comprising a local machine learning engine (LMLE, 201) comprising a set (220-222) of N artificial neural network ΑΝΝ1 , ί adapted to process N different types of input signals, the method comprising the following steps: receiving (300) a first set of N input signals S _1 (i) for authentication purpose; determining (301) respectively for each of the N input signals S _1( i ) by the N artificial neural networks ANN 1, i , N estimations LH(i) of the likelihood that a given input signal is provided by a legitimate user; determining (302) based on a risk scoring established using the N estimations LH(i) if the requesting user is authenticated as the legitimate user; if (303) the requested user is authenticated, determining (304) if at least one likelihood estimation determined for a given input signal S _1 (j ) is below a predetermined threshold T c,i, and if it is the case: transmitting (305) the input signal S _1( j ) to a remote server (210) implementing a server machine learning engine (SMLE, 211) adapted to process said N different types of input signals and trained to identify a user U_C ; receiving an input signal S _2( j ) associated to the closest candidate U_C and executing (309) by the local machine learning engine (LMLE, 201) an additional learning phase (309) using the input signal S _2( j ) as an input signal that is not associated to the requesting user.

你已经成功收藏专利！

检索式保存成功!

IPRDB

热门服务

关于我们

友情链接

联系方式