Multiple instance learning: A survey of problem characteristics and applications
作者：
Highlights：
• The characteristics specific of MIL problems are formally identified and described.
• MIL methods and applications are reviewed in the light of the problem characteristics.
• Comparative experiments show the impact of problem characteristics on 16 reference methods.
• Recommendation are issued for future benchmarking.
• Promising avenue of research are identified.
摘要
•The characteristics specific of MIL problems are formally identified and described.•MIL methods and applications are reviewed in the light of the problem characteristics.•Comparative experiments show the impact of problem characteristics on 16 reference methods.•Recommendation are issued for future benchmarking.•Promising avenue of research are identified.
论文关键词：Multiple instance learning,Weakly supervised learning,Classification,Multi-instance learning,Computer vision,Computer aided diagnosis,Document classification,Drug activity prediction
论文评审过程：Received 17 January 2017, Revised 4 August 2017, Accepted 7 October 2017, Available online 13 October 2017, Version of Record 6 February 2018.
论文官网地址：https://doi.org/10.1016/j.patcog.2017.10.009
