English

运筹学学报

• 运筹学 •    下一篇

多示例学习问题研究进展综述

田英杰1,2  胥栋宽1  张春华3,*   

  1. 1. 中国科学院虚拟经济与数据科学研究中心, 大数据挖掘与知识管理重点实验室,  北京 100190; 2. 中国科学院大学经济与管理学院，北京 100190;
    3. 中国人民大学信息学院, 北京 100872;
  • 收稿日期:2017-09-30 出版日期:2018-06-15 发布日期:2018-06-15
  • 通讯作者: 张春华 zhangchunhua@ruc.edu.cn
  • 基金资助:

    国家自然科学基金(Nos. 71731009, 61472390, 71331005, 91546201, 11771038), 北京自然科学基金(No.1162005)

A review of multi-instance learning research

TIAN Yingjie1,2   XU Dongkuan1   ZHANG Chunhua3,*   

  1. 1. Research Center on Fictitious Economy and Data Science,  the Key Laboratory of Big Data Mining and Knowledge Management, Chinese Academy of Sciences, Beijing 100190, China; 2. School of Economics and Management, University of Chinese Academy of Sciences, Beijing 100190, China; 3. School of Information, Renmin University of China, Beijing 100872, China
  • Received:2017-09-30 Online:2018-06-15 Published:2018-06-15

PDF

900

摘要：

多示例学习是一种特殊的机器学习问题，近年来得到了广泛的关注和研究，许多不同类型的多示例学习算法被提出，用以处理各个领域中的实际问题. 针对多示例学习的算法研究和应用进行了较为详细的综述, 介绍了多示例学习的各种背景假设, 从基于示例水平、包水平、嵌入空间三个方面对多示例学习的常见算法进行了描述, 并给出了多示例学习的算法拓展和若干领域的主要应用.

关键词: 多示例学习, 分类问题, 包, 支持向量机, 深度学习

Abstract:

Multi-instance learning is  a special kind of machine learning problem, has received extensive attention and been researched on in recent years. Many different types of multi-instance learning algorithms have been proposed to deal with practical problems in various fields. This paper reviews the algorithm research and application of multi-instance learning in detail, introduces various background assumptions， and introduces multi-instance learning from three aspects: instance level, bag level, and embedded space. Finally we provide the algorithm extensions  and major applications in several areas.

Key words: multi-instance learning, classification problem, bag, support vector machine, deep learning