-
公开(公告)号：US20190012580A1
公开(公告)日：2019-01-10
申请号：US16067702
申请日：2017-01-17
申请人： NEC CORPORATION
发明人： Masato ISHII
摘要： A neural network capturing a minute pattern variation useful for recognition while maintaining robustness against a pattern variation unrelated to recognition is learned. A preprocessing unit performs, on a set of patterns being to be learned and including a specific pattern variation, a plurality of preprocesses causing different degrees of the specific pattern variation. A network structure determination unit determines, for each of the plurality of preprocesses, a network structure of a neural network having robustness according to a degree of the specific pattern variation after the preprocess. A network learning unit learns, for each of the plurality of preprocesses, the neural network with the network structure associated with the preprocess using the set of patterns after the preprocess.
-
公开(公告)号：US20190012379A1
公开(公告)日：2019-01-10
申请号：US16129988
申请日：2018-09-13
申请人： NEC Corporation
发明人： Masato ISHII
摘要： Provided is a search system which is configured to search for a registered vector being similar to an input vector among a plurality of registered vectors, on the basis of a degree of similarity between an input vector and a registered vector. The search system includes a partial similarity calculation unit that calculates a degree of partial similarity which is the degree of similarity concerning some of one or more dimensions of the input vector and the registered vector, a limit calculation unit that calculates, on the basis of the degree of partial similarity, an upper limit of the degree of similarity that is expected when the degree of similarity is calculated, and a rejection decision unit that decides, on the basis of the upper limit of the degree of similarity, whether or not to reject the registered vector from a candidate for a search result.
-
公开(公告)号：US20190012378A1
公开(公告)日：2019-01-10
申请号：US16129933
申请日：2018-09-13
申请人： NEC Corporation
发明人： Masato ISHII
摘要： Provided is a search system which is configured to search for a registered vector being similar to an input vector among a plurality of registered vectors, on the basis of a degree of similarity between an input vector and a registered vector. The search system includes a partial similarity calculation unit that calculates a degree of partial similarity which is the degree of similarity concerning some of one or more dimensions of the input vector and the registered vector, a limit calculation unit that calculates, on the basis of the degree of partial similarity, an upper limit of the degree of similarity that is expected when the degree of similarity is calculated, and a rejection decision unit that decides, on the basis of the upper limit of the degree of similarity, whether or not to reject the registered vector from a candidate for a search result.
-
4.发明申请FEATURE TRANSFORMATION LEARNING DEVICE, FEATURE TRANSFORMATION LEARNING METHOD, AND PROGRAM STORAGE MEDIUM 审中-公开特征转换学习设备，特征转换学习方法和程序存储介质公开(公告)号：US20160189059A1
公开(公告)日：2016-06-30
申请号：US14909883
申请日：2014-07-25
申请人： NEC CORPORATION
发明人： Masato ISHII
CPC分类号： G06N20/00 , G06F9/4881
摘要： A feature transformation learning device includes an approximation unit, a loss calculation unit, an approximation control unit, and a loss control unit. The approximation unit takes a feature value that is extracted from a sample pattern and then weighted by a training parameter, assigns that weighted feature value to a variable of a continuous approximation function approximating a step function, and, by doing so, computes an approximated feature value. The loss calculation unit calculates a loss with respect to the task on the basis of the approximated feature value. The approximation control unit controls an approximation precision of the approximation function with respect to the step function such that the approximation function used with the approximation unit approaches the step function according to a decrease in the loss. The loss control unit updates the training parameter such that the loss decreases.
摘要翻译： 特征变换学习装置包括近似单元，损失计算单元，近似控制单元和损失控制单元。 近似单元获取从样本模式中提取的特征值，然后通过训练参数加权，将该加权特征值分配给近似步长函数的连续逼近函数的变量，并且通过这样计算近似特征 值。 损失计算单元基于近似特征值计算相对于任务的损失。 近似控制单元相对于阶梯函数控制近似函数的近似精度，使得与近似单元一起使用的近似函数根据损失的减小接近阶跃函数。 损失控制单元更新训练参数，使得损失减小。
-
公开(公告)号：US20160350416A1
公开(公告)日：2016-12-01
申请号：US15114930
申请日：2015-02-09
申请人： NEC CORPORATION
发明人： Masato ISHII
IPC分类号： G06F17/30
CPC分类号： G06F17/30743 , G06F17/30247 , G06F17/30766 , G06K9/6223
摘要： Provided is a search system which is configured to search for a registered vector being similar to an input vector among a plurality of registered vectors, on the basis of a degree of similarity between an input vector and a registered vector. The search system includes a partial similarity calculation unit that calculates a degree of partial similarity which is the degree of similarity concerning some of one or more dimensions of the input vector and the registered vector, a limit calculation unit that calculates, on the basis of the degree of partial similarity, an upper limit of the degree of similarity that is expected when the degree of similarity is calculated, and a rejection decision unit that decides, on the basis of the upper limit of the degree of similarity, whether or not to reject the registered vector from a candidate for a search result.
摘要翻译： 提供了一种搜索系统，其被配置为基于输入向量和注册向量之间的相似度来搜索与多个注册向量中的输入向量相似的注册向量。 搜索系统包括部分相似度计算单元，其计算作为与输入向量和登记向量的一个或多个维度中的一些相关的相似度的部分相似度的程度;限制计算单元，基于 计算部分相似度的程度，计算相似度时期望的相似程度的上限;以及拒绝判定单元，根据相似度的上限判定是否拒绝 来自搜索结果候选者的注册向量。
-
公开(公告)号：US20210216563A1
公开(公告)日：2021-07-15
申请号：US15734442
申请日：2018-06-04
申请人： NEC Corporation
发明人： Devendra DHAKA , Masato ISHII , Atsushi SATO
IPC分类号： G06F16/25 , G06F16/28 , G06F16/2455 , G06F16/21
摘要： The information processing apparatus (2000) of the example embodiment 1 includes an acquisition unit (2020), a clustering unit (2040), a transformation unit (2060) and modeling unit (2080). Until a predetermined termination condition is determined, the clustering unit (2040) repeatedly preforms: 1) optimizing the posterior parameters for clustering assignment for each data streams; 2) optimizes the posterior parameters for each determined cluster and for each time frame; 3) optimizes the posterior parameters for individual responses for each data stream; 4) optimizes the posterior parameters for latent states, via approximating the observation model through non-conjugate inference. The transformation unit (2060) transforms the latent states into parameters of the observation model, through a transformation function. The modeling unit (2060) generates the model data, which including all the optimized parameters of all the model latent variables, optimized inside the clustering unit (2040).
-
公开(公告)号：US20190012380A1
公开(公告)日：2019-01-10
申请号：US16129994
申请日：2018-09-13
申请人： NEC Corporation
发明人： Masato ISHII
摘要： Provided is a search system which is configured to search for a registered vector being similar to an input vector among a plurality of registered vectors, on the basis of a degree of similarity between an input vector and a registered vector. The search system includes a partial similarity calculation unit that calculates a degree of partial similarity which is the degree of similarity concerning some of one or more dimensions of the input vector and the registered vector, a limit calculation unit that calculates, on the basis of the degree of partial similarity, an upper limit of the degree of similarity that is expected when the degree of similarity is calculated, and a rejection decision unit that decides, on the basis of the upper limit of the degree of similarity, whether or not to reject the registered vector from a candidate for a search result.
-
8.发明申请公开(公告)号：US20180018538A1
公开(公告)日：2018-01-18
申请号：US15545769
申请日：2016-02-05
申请人： NEC Corporation
发明人： Masato ISHII
CPC分类号： G06K9/6215 , G06K9/00624 , G06K9/62 , G06K9/6247 , G06K9/6279 , G06K9/68 , G06K9/6807 , G06N20/00 , G06N20/10 , G06N99/00
摘要： Provided are a feature transformation device and others enabling feature transformation with high precision.The feature transformation device includes receiving means for receiving training data and test data each including a plurality of samples, optimization means for optimizing weight and feature transformation parameter based on an objective function related to the weight and the feature transformation parameter, the optimization means including weight derivation means for deriving the weight assigned to each element included in the training data and feature transformation parameter derivation means for deriving the feature transformation parameter that transforms each of the samples included in the training data or the test data, objective function derivation means for deriving a value of the objective function, the objective function derivation means including a constraint determination means for determining whether the weight satisfies a prescribed constraint and regularization means for regularizing at least one of the weight or the feature transformation parameter, and transformation means for transforming an element included in at least one of the training data or the test data based on the feature transformation parameter.
-
公开(公告)号：US20220245518A1
公开(公告)日：2022-08-04
申请号：US17610488
申请日：2019-05-22
申请人： NEC Corporation
摘要： A data transformation apparatus (1) includes: data transformation means (11) for performing data transformation on each of a plurality of data sets so that data distributions of the plurality of data sets are brought close to each other; first calculation means (12) for calculating a class classification loss from a result of class classification performed by class classification means on at least some of a plurality of first transformed data sets obtained after the data transformation; second calculation means (13) for calculating an upper bound and a lower bound of a domain classification loss from a result of domain classification performed by domain classification means on each of the plurality of first transformed data sets; and first learning means (14) for performing first learning by updating a parameter of the domain classification means so that the upper bound is reduced and updating a parameter of the data transformation means so that the class classification loss is reduced and the lower bound is increased.
-
公开(公告)号：US20220121990A1
公开(公告)日：2022-04-21
申请号：US17422678
申请日：2019-01-22
申请人： NEC Corporation
摘要： A data conversion learning apparatus includes a data conversion unit that performs data conversion of source data and target data, a first deduction unit that deduces data of a non-appearing class on the basis of a domain certainty factor acquired by a domain identification using converted data, a second deduction unit that deduces data of a non-appearing class on the basis of a class certainty factor acquired by a class identification using converted data, a class identification learning unit that performs machine learning for class identification using the data of the non-appearing class deduced by the first deduction unit and the source data and the target data which are inputs, and a domain identification learning unit that performs machine learning for domain identification using the data of the non-appearing class deduced by the second deduction unit and the source data and the target data which are inputs.
-
-
-
-
-
-
-
-
-
搜索结果
17 条记录 (耗时 0.015 秒)
