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. 发明授权

EP2570959B1 A method of estimating the potential damage of data leakage incident by assigning misuseability weight 有权转让
公开(公告)号：EP2570959B1
公开(公告)日：2018-07-18
申请号：EP12006214.6
申请日：2012-09-03
申请人： Deutsche Telekom AG
发明人： Harel, Amir , Shabtai, Asaf , Rokach, Lior , Elovici, Yuval
IPC分类号： G06F21/62
CPC分类号： G06F21/6254 , G06F21/6227
摘要： A computerized method for detecting data leakage and data misuse in a computer system belonging to an organization. The method comprising the steps of determining a raw record score; determining a record distinguishing factor; determining a final record score using the distinguishing factor and the raw record score; and combining the final record score with a quantity factor in order to determine the misuseability weight.

2. 发明公开

EP2570959A2 A method of estimating the potential damage of data leakage incident by assigning misuseability weight 有权转让
标题翻译：通过分配加权估算数据丢失事件滥用的潜在损伤的方法
公开(公告)号：EP2570959A2
公开(公告)日：2013-03-20
申请号：EP12006214.6
申请日：2012-09-03
申请人： Deutsche Telekom AG
发明人： Harel, Amir , Shabtai, Asaf , Rokach, Lior , Elovici, Yuval
IPC分类号： G06F21/50 , G06F21/62
CPC分类号： G06F21/6254 , G06F21/6227
摘要： A computerized method for detecting data leakage and data misuse in a computer system belonging to an organization. The method comprising the steps of determining a raw record score; determining a record distinguishing factor; determining a final record score using the distinguishing factor and the raw record score; and combining the final record score with a quantity factor in order to determine the misuseability weight.
摘要翻译：用于在组织属于计算机系统中检测数据泄漏和数据滥用的计算机化的方法。该方法包括确定开采原始分数记录的步骤; 确定一个创纪录的区别因子; 确定性开采使用区别因子和原始分数记录最终的得分纪录; 并结合最终的得分纪录与数量因素，以确定矿misuseability重量。

3. 发明公开

EP2079040A1 An improved stacking schema for classification tasks 审中-公开
标题翻译： Verbessertes StapelschemafürKlassifizierungsaufgaben
公开(公告)号：EP2079040A1
公开(公告)日：2009-07-15
申请号：EP09000173.6
申请日：2009-01-08
申请人： Deutsche Telekom AG
发明人： Menahem, Eitan , Rokach, Lior , Elovici, Yuval
IPC分类号： G06K9/68
CPC分类号： G06K9/6292 , G06N5/043 , G06N99/005
摘要： a method for improving stacking schema for classification tasks, according to which predictive models are built, based on stacked-generalization meta-classifiers. Classifications are combined to build a new scheme from at least two layers and multiclass classification problems are converted into binary classification problems. One-against-all class binarization and regression learners are used for each class model and ensemble classifiers are improved using stacking. Accuracy differences, accuracy ratio, and runtime classification in multiclass datasets are also improved and the class of a value is then predicted.
摘要翻译：一种用于改进分类任务的堆叠模式的方法，根据该方法构建基于堆叠泛化元分类器的预测模型。将分类结合起来，从至少两层构建一个新的方案，将多类分类问题转化为二进制分类问题。对于每个类模型使用一对二类二值化和回归学习者，并且使用堆叠来改进整体分类器。多类数据集中的精度差异，准确率和运行时间分类也得到改善，然后预测值的类别。

4. 发明公开

EP3091497A1 RECOMMENDATION SYSTEM AND METHOD FOR A MOBILE DEVICE BASED ON RAW DATA WHICH IS COLLECTED FROM SENSORS OF THE MOBILE DEVICE 审中-公开
标题翻译：用于在移动设备基传感器从所述移动设备建议的系统和方法收集的原始数据
公开(公告)号：EP3091497A1
公开(公告)日：2016-11-09
申请号：EP16166381.0
申请日：2016-04-21
申请人： Deutsche Telekom AG
发明人： Shapira, Bracha , Rokach, Lior , Bar, Ariel , Unger, Moshe , Chizi, Barak
IPC分类号： G06Q30/06 , G06Q30/02
CPC分类号： G06Q30/0631 , G06Q30/0267
摘要： The present invention refers to a recommendation system which comprises:
e. extracting unit at a user mobile device for extracting latent raw sensors data from one or more sensors of the mobile device;
f. communication means for conveying said latent raw sensors data, optionally together with passive data, to a query formation module at a remote server;
g. said query formation module at the remote server for forming a query from said latent raw sensors data, and optionally also from said passive data and from a user profile data, if exists within said query formation module, and conveying the query into a recommender at said remote server;
h. said recommender at the remote location for: (i) receiving said query from said query formation module; (ii) processing said query against a general model and a recommendation library, to obtain a product recommendation R; and (iii) conveying said product recommendation R to the user mobile device.
摘要翻译：本发明涉及一种推荐系统，该系统包括：e。在用于提取从所述移动设备的一个或多个传感器潜原始传感器数据的用户移动装置提取单元; F。通信装置，用于传送所述潜原始传感器数据，具有被动数据任选地，在一个远程服务器的查询形成模块; 克所述用于从所述被动数据和从用户简档数据从所述形成的查询潜原始传感器数据，因此可选地，如果所述查询形成模块内时，和传送查询分成在所述一个推荐的远程服务器查询形成模块 REMOTESERVER; 小时。所述推荐器在远程位置：（i）从查询接收所述查询，所述形成模块; （ii）加工所述查询针对一般模型和推荐文库，以获得产物推荐R等和（iii）传送所述产品推荐R键的用户移动装置。

5. 发明公开

EP2960849A1 METHOD AND SYSTEM FOR RECOMMENDING AN ITEM TO A USER 审中-公开
标题翻译： VERFAHREN UND SYSTEM ZUM EMPFEHLEN EINES GEGENSTANDES AN EINEN BENUTZER
公开(公告)号：EP2960849A1
公开(公告)日：2015-12-30
申请号：EP15174025.5
申请日：2015-06-26
申请人： Deutsche Telekom AG
发明人： Grolman, Edita , Bar, Ariel , Shapira, Bracha , Rokach, Lior , Dayan, Aviram
IPC分类号： G06Q30/06 , G06Q30/02
CPC分类号： G06Q30/0631 , G06Q30/0282
摘要： A computer implemented method for recommending an item to a user, according to which a plurality of users each of which with a known profile, interface with a corresponding computerized device and binary event-specific, user-item source matrices are generated, for indicating whether a given user performed a given explanatory event included in a corresponding source matrix. All users included in the source matrices are grouped to a profile-specific cluster and items included in the source matrices are grouped to an item category cluster to generate a predictive event book matrix which indicates the probability that an event unknown to have been performed by the given user will be performed for each profile-specific and item category cluster combination included in the one or more source matrices. Each user gets recommendation about an item included in the source matrices, which has a highest probability.
摘要翻译：一种用于向用户推荐项目的计算机实现的方法，根据该方法，生成具有已知简档的多个用户与对应的计算机化设备和二进制事件特定的用户项目源矩阵的接口，用于指示是否给定用户执行包括在相应源矩阵中的给定说明事件。包括在源矩阵中的所有用户被分组到特定于配置文件的集群，并且包括在源矩阵中的项目被分组到项目类别集群以生成预测事件簿矩阵，其指示事件未被执行的事件的概率给定用户将针对包含在一个或多个源矩阵中的每个特定于个人资料和项目类别集群组合执行。每个用户获得关于源矩阵中包括的具有最高概率的项目的推荐。

6. 发明公开

EP2860672A3 Scalable cross domain recommendation system 审中-公开
标题翻译：可扩展的，跨领域推荐系统
公开(公告)号：EP2860672A3
公开(公告)日：2015-08-12
申请号：EP14188329.8
申请日：2014-10-09
申请人： Deutsche Telekom AG
发明人： Shapira, Bracha , Bar, Ariel , Grolman, Edita , Dayan, Aviram , Rokach, Lior
IPC分类号： G06N99/00
CPC分类号： G06N99/005
摘要： A method for generating recommendation for a target domain based on data from other source domains is provided, comprising the steps of: aggregating data in each of said other domains; splitting said aggregated data to a train data set, a validation data set and a test data set; extracting patterns in each of said source domains; transferring said extracted patterns from said source domains to said target domain while testing different models and learning the best model; and generating recommendation for said target domain according to said learned best model.

7. 发明公开

EP2490149A1 System for verifying user identity via mouse dynamics 审中-公开
标题翻译： System zurÜberprüfungderBenutzeridentitätüberMausdynamik
公开(公告)号：EP2490149A1
公开(公告)日：2012-08-22
申请号：EP12000611.9
申请日：2012-01-31
申请人： Deutsche Telekom AG
发明人： Feher, Clint , Moskovitch, Rober , Rokach, Lior , Elovici, Yuval
IPC分类号： G06F21/20
CPC分类号： G06F21/316
摘要： The invention relates to a system which accumulates during a training phase features of users' behaviors that are characteristic for each specific user while operating its pointing device. Thereafter, the system monitors the operation of the pointing devices, and identifies situations when one or more pointing devices are operated by users that have not participated in said training phase. The system alerts upon such identified situations.
摘要翻译：本发明涉及一种系统，其在训练阶段期间积累用户行为的特征，这些特征在每个特定用户操作其指点设备时都是特征的。此后，系统监视指示设备的操作，并且识别当没有参与所述训练阶段的用户操作一个或多个指点设备时的情况。系统会根据这些识别的情况进行警报。

8. 发明授权

EP2477136B1 Method for continuously verifying user identity via keystroke dynamics 有权转让
标题翻译：通过按键动态连续验证用户身份的方法
公开(公告)号：EP2477136B1
公开(公告)日：2018-04-11
申请号：EP12000082.3
申请日：2012-01-09
申请人： Deutsche Telekom AG
发明人： Shimshon, Tomer , Moskovitch, Robert , Rokach, Lior , Elovici, Yuval
IPC分类号： G06F21/30
CPC分类号： G06F21/316 , G06F21/32
摘要： A method of verifying the identity of users according to their interaction with a keyboard device of a computer, according to which during a training phase in which a group of users are trained, input streams of keystrokes performed by each of the trained users are recorded and similar pairs of consecutive keystrokes are joined, based on their temporal features. A features vector is then extracted from each input stream of keystrokes, so as to create a user verification model representing the user typing patterns. During a verification phase in which real-time streams of keystrokes are verified, input streams of keystrokes performed by a verified user are recorded and each of the streams is transformed to a features vector. A vector of probabilities indicating the probabilities that the features vector belongs to each user is generated and the stream of keystrokes and associating it with the user with the highest probability is classified.

9. 发明公开

EP3179434A1 DESIGNING CONTEXT-AWARE RECOMMENDATION SYSTEMS, BASED ON LATENT CONTEXTS 审中-公开
标题翻译：基于潜在的上下文设计上下文感知的推荐系统
公开(公告)号：EP3179434A1
公开(公告)日：2017-06-14
申请号：EP16203148.8
申请日：2016-12-09
申请人： Deutsche Telekom AG
发明人： Shapira, Bracha , Rokach, Lior , Bar, Ariel , Unger, Moshe , Chizi, Barak
IPC分类号： G06Q30/02 , G06F17/30
CPC分类号： G06Q30/0631
摘要： A system for generating a context-aware recommendation for a target user and a target item based on latent context, which comprises a data collection unit configured to collect raw sensor data; a sensor data feature extractor, configured to extract features from the raw sensor data; a latent content extractor, configured to extract latent context from the feature; a data matcher, configured to merge all the collected information to create a rating model; a latent context matrix factorization component, configured to receive as an input the rating model schema and generate a recommendation model; and a Model evaluator, configured to evaluate the generated model accuracy.
摘要翻译：一种用于基于潜在上下文产生针对目标用户和目标项目的上下文感知推荐的系统，其包括配置成收集原始传感器数据的数据收集单元; 传感器数据特征提取器，被配置为从原始传感器数据中提取特征; 潜在内容提取器，被配置为从所述特征提取潜在上下文; 数据匹配器，被配置为合并所有收集的信息以创建评级模型; 潜在上下文矩阵分解组件，被配置为接收评级模型模式作为输入并生成推荐模型; 和模型评估器，用于评估生成的模型精度。

10. 发明公开

EP2983123A1 SELF TRANSFER LEARNING RECOMMENDATION METHOD AND SYSTEM 审中-公开
标题翻译： EMPFEHLUNGSVERFAHREN UND -SYSTEMFÜRSELBSTÜBERTRAGENDESLERNEN
公开(公告)号：EP2983123A1
公开(公告)日：2016-02-10
申请号：EP15179606.7
申请日：2015-08-04
申请人： Deutsche Telekom AG
发明人： Grolman, Edita , Iny, Yoni , Bar, Ariel , Shapira, Bracha , Rokach, Lior , Dayan, Aviram
IPC分类号： G06Q30/02 , G06Q30/00
CPC分类号： G06Q30/00 , G06Q30/02
摘要： A computer implemented method for recommending an item to a user in big data systems, according to which a binary matrix of ratings of items is generated by users for a full target domain and sub-matrixes with high correlation between the users, items and their ratings are discovered in the binary matrix. The density of the sub-matrixes in the binary matrix is calculated and a source domain matrix is populated according to the results, based on the density calculated in the discovered sub-matrixes. Then top-N matrixes with highest densities are selected. Codebooks, which contains different user-item rating patterns, are constructed for each of the selected sub-matrixes, which are sub-domains and constructing a codebook for the matrix with the full rating dataset. Finally the captured different user-item rating patterns are projected to the full domain dataset, based on the constructed codebooks.
摘要翻译：一种计算机实现的方法，用于向大数据系统中的用户推荐项目，根据该方法，用户为完整目标域生成项目评级的二进制矩阵，并且在用户，项目及其评级之间具有高相关性的子矩阵在二进制矩阵中被发现。计算二进制矩阵中子矩阵的密度，并根据结果，根据发现的子矩阵中计算的密度填充源域矩阵。然后选择具有最高密度的top-N矩阵。为每个选定的子矩阵构建包含不同用户项目评估模式的代码簿，这些子矩阵是子域，并使用完整的分级数据集构建矩阵的码本。最后，基于构建的码本，将捕获的不同用户项目评级模式投影到完整域数据集。

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