Multi-label active learning through serial–parallel neural networks
作者：
Highlights：
• We propose multi-label active learning through serial-parallel neural networks.
• Serial and parallel parts serve for feature extraction and label prediction.
• Pairwise outputs support missing label handling and label uncertainty calculation.
• Label query integrates instance representativeness, label uncertainty and scarcity.
• Results validate the effectiveness of MASP compared to three sets of algorithms.
摘要
•We propose multi-label active learning through serial-parallel neural networks.•Serial and parallel parts serve for feature extraction and label prediction.•Pairwise outputs support missing label handling and label uncertainty calculation.•Label query integrates instance representativeness, label uncertainty and scarcity.•Results validate the effectiveness of MASP compared to three sets of algorithms.
论文关键词：Label correlation,Missing label,Multi-label active learning,Neural network,Query strategy
论文评审过程：Received 20 February 2022, Revised 27 May 2022, Accepted 7 June 2022, Available online 15 June 2022, Version of Record 23 June 2022.
论文官网地址：https://doi.org/10.1016/j.knosys.2022.109226
