Multi-instance multi-label active learning
Multi-instance multi-label active learning
复制标题
DOI：
10.24963/ijcai.2017/262
复制
发表时间：
2017-08
期刊：
影响因子：
--
通讯作者：
Sheng-Jun Huang;Nengneng Gao;Songcan Chen
中科院分区：
文献类型：
--
作者：
Sheng-Jun Huang;Nengneng Gao;Songcan Chen
多示例多标记学习（MIML）在各种应用中取得了成功，特别是涉及复杂学习对象的应用。沿着表达能力的增强，标注MIML示例的开销也显著增加。在本文中，我们提出了一种新的主动学习方法，以减少标记成本的MIML。该方法通过利用输入和输出空间中的多样性和不确定性来主动查询最有价值的信息。它专门针对MIML对象设计了一种新颖的查询策略，并在不增加额外成本的情况下从Oracle获取更精确的信息。基于查询的信息，MIML模型，然后有效地训练，同时优化实例和标签之间的相关性排名。在基准数据集上的实验表明，该方法在各种标准下都取得了上级性能。
Multi-instance multi-label learning (MIML) has achieved success in various applications, especially those involving complicated learning objects. Along with the enhancing of expressive power, the cost of annotating a MIML example also increases significantly. In this paper, we propose a novel active learning approach to reduce the labeling cost of MIML. The approach actively query the most valuable information by exploiting diversity and uncertainty in both the input and output spaces. It designs a novel query strategy for MIML objects specifically and acquires more precise information from the oracle without additional cost. Based on the queried information, the MIML model is then effectively trained by simultaneously optimizing the relevance rank among instances and labels. Experiments on benchmark datasets demonstrate that the proposed approach achieves superior performance on various criteria.