公开(公告)号：US09928448B1

公开(公告)日：2018-03-27

申请号：US15273872

申请日：2016-09-23

Applicant: International Business Machines Corporation

Inventor： Michele Merler ,  John R. Smith ,  Rosario A. Uceda-Sosa ,  Hui Wu

IPC: G06K9/00 ,  G06K9/66 ,  G06K9/62

CPC classification number: G06K9/66 ,  G06K9/627 ,  G06K9/6277 ,  G06K9/628 ,  G06K9/6285 ,  H05K999/99

Abstract: A method includes utilizing two or more classifiers to calculate, for an input image, probability scores for a plurality of classes based on visual information extracted from the input image and semantic relationships in a classification hierarchy, wherein each of the two or more classifiers is associated with a given one of two or more levels in the classification hierarchy with each level in the classification hierarchy comprising a subset of the plurality of classes, and classifying the input image based on the calculated probability scores.

公开(公告)号：US09477973B2

公开(公告)日：2016-10-25

申请号：US13926544

申请日：2013-06-25

Applicant: International Business Machines Corporation

Inventor： Bernadette R. Duponchel ,  Myron Flickner ,  Victoria B. Platt ,  John R. Smith

IPC: G06F3/00 ,  G06F3/048 ,  G06Q30/02

CPC classification number: G06Q30/0643 ,  G06Q30/0255 ,  G06Q30/0269 ,  G06Q30/0277 ,  G06Q30/0282 ,  G06Q50/01 ,  G06T11/001 ,  H04L67/306

Abstract: A personalization enhancement method and system. The method includes retrieving and analyzing digital content associated with a user. Characteristics describing the digital content are tagged and transferred to a profile of said the user. The profile includes additional characteristics generated during previous analysis of data from the digital content and additional digital content associated with the user. User information associated with products, a location, and a time profile is assigned. The profile is analyzed based on selection and interaction of the user with respect to a consumer Website. The profile includes the characteristics and the additional characteristics with respect to products and services of the consumer Website. A presentation color setting and a group of products and services of are determined for presentation to the user. The group of products and services are presented to the user using the presentation color setting.

Abstract translation: 个性化增强方法和系统。 该方法包括检索和分析与用户相关联的数字内容。 描述数字内容的特征被标记并传送到所述用户的简档。 该简档包括在先前分析来自数字内容的数据和与用户相关联的附加数字内容期间生成的附加特征。 分配与产品相关联的用户信息，位置和时间简档。 基于用户对消费者网站的选择和交互来分析简档。 该简介包括消费者网站的产品和服务的特征和附加特征。 确定呈现颜色设置和一组产品和服务以呈现给用户。 使用演示颜色设置向用户显示一组产品和服务。

公开(公告)号：US20160092789A1

公开(公告)日：2016-03-31

申请号：US14500023

申请日：2014-09-29

Applicant: International Business Machines Corporation

Inventor： Noel C. Codella ,  Gang ` Hua ,  John R. Smith

IPC: G06N99/00

CPC classification number: G06N20/00

Abstract: Methods, systems, and computer program products for category oversampling for imbalanced machine learning are provided herein. A method includes identifying an anchor data point in a given class of data points underrepresented among multiple classes in a data set of multiple data points, wherein each data point represent a vector; determining a number of data points in the given class that neighbor the anchor data point, wherein the number comprises two or more; applying a weight to (i) each of the number of data points to create a number of weighted neighboring data points, and (ii) the anchor data point to create a weighted anchor data point, wherein the sum of all weights is equal to one; performing a vector summation by summing the number of weighted neighboring data points and the weighted anchor data point; and generating a synthetic data point based on said vector summation.

Abstract translation: 本文提供了用于不平衡机器学习的类别过采样的方法，系统和计算机程序产品。 一种方法包括识别在多个数据点的数据集中的多个类中表示的给定类别的数据点中的锚点数据点，其中每个数据点表示向量; 确定所述给定类中与所述锚点数据点相邻的多个数据点，其中所述数目包括两个或更多个; 对（i）多个数据点中的每一个施加权重以创建多个加权相邻数据点，以及（ii）所述锚数据点以创建加权的锚数据点，其中所有权重的和等于一 ; 通过加权相邻数据点的数量和加权的锚点数据点相加来执行向量求和; 以及基于所述向量求和生成合成数据点。

公开(公告)号：US20140205186A1

公开(公告)日：2014-07-24

申请号：US13744688

申请日：2013-01-18

Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION

Inventor： Liangliang Cao ,  Noel C.F. Codella ,  Gang Hua ,  John R. Smith

IPC: G06K9/62

CPC classification number: G06K9/6256 ,  G06K9/00657

Abstract: Techniques for generating cross-modality semantic classifiers and using those cross-modality semantic classifiers for ground level photo geo-location using digital elevation are provided. In one aspect, a method for generating cross-modality semantic classifiers is provided. The method includes the steps of: (a) using Geographic Information Service (GIS) data to label satellite images; (b) using the satellite images labeled with the GIS data as training data to generate semantic classifiers for a satellite modality; (c) using the GIS data to label Global Positioning System (GPS) tagged ground level photos; (d) using the GPS tagged ground level photos labeled with the GIS data as training data to generate semantic classifiers for a ground level photo modality, wherein the semantic classifiers for the satellite modality and the ground level photo modality are the cross-modality semantic classifiers.

公开(公告)号：US20140085328A1

公开(公告)日：2014-03-27

申请号：US13625132

申请日：2012-09-24

Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION

Inventor： Noel C. Codella ,  Apostol Natsev ,  John R. Smith

IPC: G06K9/46 ,  G09G5/377 ,  G06K9/62

CPC classification number: G06N5/04 ,  G06F17/30 ,  G06F17/30247 ,  G06F17/30265 ,  G06F17/30598 ,  G06K9/6267 ,  G06N7/005 ,  G06N99/005

Abstract: Techniques for detecting an event via social media content. A method includes obtaining multiple images from at least one social media source, extracting at least one visual semantic concept from the multiple images, differentiating an event semantic concept signal from a background semantic concept signal to detect an event in the multiple images, and retrieving one or more images associated with the event semantic concept signal for presentation as a visual description of the detected event.

Abstract translation: 通过社交媒体内容检测事件的技术。 一种方法包括从至少一个社交媒体源获得多个图像，从多个图像中提取至少一个视觉语义概念，将事件语义概念信号与背景语义概念信号区分开来，以检测多个图像中的事件，并且检索一个 或更多与事件语义概念信号相关联的图像，以作为所检测到的事件的视觉描述。