-
公开(公告)号：EP3196821A4
公开(公告)日：2018-01-24
申请号：EP15838688
申请日：2015-06-22
Applicant: SONY CORP
Inventor： YAMAMOTO NORIYUKI , TAKAMURA SEIICHI , SEKINO MASASHI
CPC classification number: G06F17/30123 , G06Q30/0202 , G06Q30/0252 , G06Q30/0631
Abstract: To make it possible to enhance convenience for users and a seller of content. There is provided an information processing device including: a schema converting unit that converts content meta-information managed by a plurality of different management systems into a common schema; and a recommendation unit that determines a combination of content to be recommended and a user based on commonized content meta-information which is obtained by the conversion into the common schema and content purchase history information in the respective management systems.
-
Patent Agency Ranking2.发明公开INFORMATION PROCESSING DEVICE, INFORMATION PROCESSING METHOD, AND PROGRAM 审中-公开INFORMATIONSVERARBEITUNGSVORTHHTUNG，INFORMATIONSVERARBEITUNGSVERFAHREN UND PROGRAMM
公开(公告)号：EP3029620A4
公开(公告)日：2017-04-12
申请号：EP14831723
申请日：2014-07-22
Applicant: SONY CORP
Inventor： YAMAMOTO NORIYUKI , KASAMI YUTAKA , TAKAMURA SEIICHI , ASANO YASUHARU , SEKINO MASASHI
CPC classification number: G06Q10/02 , G06F17/30867 , G06Q30/02 , G06Q30/0631 , G06Q50/10 , G06Q50/14
-
3.发明专利
公开(公告)号：JP2013024792A
公开(公告)日：2013-02-04
申请号：JP2011161758
申请日：2011-07-25
Inventor： SEKINO MASASHI , KATO YASUNOBU , ITO TATSUMI
IPC: G01N21/64
CPC classification number: G01N21/6428 , G01N21/6408 , G01N21/6458
Abstract: PROBLEM TO BE SOLVED: To more accurately calculate a true spectral intensity on the basis of a measured spectrum.SOLUTION: Provided is an information processing apparatus including an estimation unit that expresses a light intensity distribution, which is obtained by irradiating light to a measurement object of a measurement target having a plurality of substances with mutually different responsive characteristics to the light on a surface of and/or inside it, as a linear combination of light intensity distributions obtained by irradiating the light to reference measurement objects, each of which has only one of the plurality of substances, models the light intensity distributions obtained from the respective reference measurement objects on the assumption that they obey predetermined probability distributions, and estimates combination coefficients of the linear combination from the light intensity distribution obtained from the measurement object of the measurement target.
Abstract translation: 要解决的问题：基于测量的光谱更准确地计算真实光谱强度。 解决方案：提供了一种信息处理设备，包括：表示光强度分布的估计单元，该估计单元通过向具有相互不同的响应特性的多个物质的测量对象的光照射到光上而获得 作为通过将光照射到参考测量对象而获得的光强度分布的线性组合，每个参考测量对象仅具有多个物质中的一个，对从各个参考测量获得的光强度分布进行建模 假设它们服从预定概率分布，并且根据从测量对象的测量对象获得的光强度分布来估计线性组合的组合系数。 版权所有（C）2013，JPO＆INPIT
-
4.发明专利
公开(公告)号：JP2013246140A
公开(公告)日：2013-12-09
申请号：JP2012122161
申请日：2012-05-29
Inventor： SEKINO MASASHI , KATO YASUNOBU , ITO TATSUMI
CPC classification number: G01N21/6486 , G01N15/1429 , G01N2015/1477 , G01N2021/6441 , G06F19/10
Abstract: PROBLEM TO BE SOLVED: To figure out more exactly the true spectral intensity on the basis of a measured spectrum.SOLUTION: An information processing apparatus comprises: an assaying unit that statistically assays each frequency band by using, as objects of comparison, simple-stained data obtained by fluorescence measuring of particles simply stained with a staining substance having prescribed fluorescence characteristics and non-stained data obtained by fluorescence measuring of unstained particles; a masking unit that sets, for each frequency band, in the absence of any significant difference between the simple-stained data and the non-stained data, the simple-stained data to 0 or a prescribed value; and an estimating unit that estimates, on the basis of the assumption that double-stained data obtained by fluorescence measuring of particles stained with a plurality of staining substances is expressed by a linear combination of base vectors representing the distribution of pieces of simple-stained data each corresponding to each staining substance, the combination coefficient of the pertinent linear combination.
Abstract translation: 要解决的问题：基于测量的光谱更准确地确定真实的光谱强度。解决方案：信息处理设备包括：测定单元，其通过使用作为比较对象的简单染色来统计分析每个频带 通过用具有规定荧光特性的染色物质简单染色的颗粒的荧光测量获得的数据和通过未染色颗粒的荧光测量获得的非染色数据获得的数据; 对于每个频带，在简单染色数据和非染色数据之间没有任何显着差异的情况下，将简单染色数据设置为0或规定值，设置掩蔽单元; 以及估计单元，其基于如下假定：通过用多种染色物质染色的颗粒的荧光测量获得的双重染色数据由表示简单染色数据片段分布的基本矢量的线性组合来表示 每个对应于每个染色物质，相关线性组合的组合系数。
-
5.发明专利
公开(公告)号：JP2012058972A
公开(公告)日：2012-03-22
申请号：JP2010200980
申请日：2010-09-08
Inventor： SEKINO MASASHI
CPC classification number: G06Q30/0282 , G06Q10/101
Abstract: PROBLEM TO BE SOLVED: To provide an evaluation prediction device capable of predicting an evaluation value at high speed.SOLUTION: An evaluation prediction device comprises: a posterior distribution calculation part for calculating variational posterior distributions of a first potential vector showing a potential feature of a first item and a second potential vector showing a potential feature of a second item, by regarding each residual matrix Rh of rank h (h=0 to H) in an evaluation value matrix of a rank quantity H that, as factor, has an evaluation value expressed by the first potential vector, the second potential vector and an inner product of the first and second potential vectors, as a random variable following a normal distribution, and by executing a variable Bayesian estimate using a known evaluation value provided as learning data; and an evaluation value prediction part for predicting the evaluation value that is unknown yet, using the variable posterior distribution of the first and second potential vectors that is calculated by the posterior distribution calculation part.
Abstract translation: 要解决的问题：提供能够高速地预测评价值的评价预测装置。 解决方案：评估预测装置包括：后验分布计算部分，用于计算示出第一项目的潜在特征的第一潜在向量的变分后验分布，以及示出第二项目的潜在特征的第二潜在矢量，通过关于 作为因子具有由第一电位矢量表示的评估值，第二电位矢量和第二电位矢量的内积的秩H的评价值矩阵中的h（h = 0〜H）的残差矩阵Rh 第一和第二电位矢量，作为正态分布之后的随机变量，并且使用提供为学习数据的已知评估值执行可变贝叶斯估计; 以及评估值预测部，用于使用由所述后验分布计算部计算出的所述第一和第二电位矢量的所述可变后验分布来预测未知的所述评价值。 版权所有（C）2012，JPO＆INPIT
-
公开(公告)号：JP2011248829A
公开(公告)日：2011-12-08
申请号：JP2010124165
申请日：2010-05-31
Inventor： SEKINO MASASHI
IPC: G06F17/30
CPC classification number: G06F17/3053 , G06F17/30867 , G06N99/005
Abstract: PROBLEM TO BE SOLVED: To further enhance recommendation precision.SOLUTION: An evaluation prediction device is provided which defines first and second latent vectors respectively corresponding to first and second items, an evaluation value expressed by an internal product of the first and the second latent vectors corresponding to each of the combinations of the first and the second items, first and second characterizing vectors respectively corresponding to the first and the second items, first and second projection matrix for projecting the first and the second characterizing vectors in space of the first and the second latent vectors. According to a normal distribution with the projection value of the first and the second characterizing vectors through the first and the second projection matrix as an expected value, the first and the second latent vectors are expressed. A Bayesian estimation is performed by using the first characterizing vector, the second characterizing vector and an existing evaluation value as study data, thereby calculating the posterior distribution of the parameter groups. Based on the posterior distribution of the parameter groups, the distribution of unknown evaluation values is calculated.
Abstract translation: 要解决的问题：进一步提高推荐精度。 解决方案：提供一种评估预测装置，其分别对应于第一和第二项目的第一和第二潜在向量，由与第一和第二项目的组合对应的第一和第二潜在向量的内积表示的评价值 第一和第二项目，分别对应于第一和第二项目的第一和第二特征向量，用于在第一和第二潜在向量的空间中投影第一和第二特征向量的第一和第二投影矩阵。 根据通过第一和第二投影矩阵的第一和第二特征向量的投影值作为期望值的正态分布，第一和第二潜在向量被表示。 通过使用第一特征向量，第二特征向量和现有评价值作为研究数据进行贝叶斯估计，从而计算参数组的后验分布。 根据参数组的后验分布，计算未知评估值的分布。 版权所有（C）2012，JPO＆INPIT
-
-
-
-
-
Search Results
6
records
(took 0.026 seconds)
