A Sequential Multi-task Learning Neural Network with Metric-Based Knowledge Transfer
A Sequential Multi-task Learning Neural Network with Metric-Based Knowledge Transfer
复制标题
具有基于度量的知识转移的顺序多任务学习神经网络
DOI：
10.1109/icmla.2012.125
复制
发表时间：
2012
期刊：
影响因子：
--
通讯作者：
S. Yue and S. Ozawa
中科院分区：
文献类型：
--
作者：
A. A. Joseph;Y.-M. Jang;S. Ozawa;and M. Lee;Tetsuyuki Takahama,Setsuko Sakai;S. Yue and S. Ozawa
在本文中，我们提出了一个新的顺序多任务模式识别模型称为资源分配网络多任务学习与度量学习（RAN-MTLML）。RAN-MTLML具有以下五个功能：一次性增量学习，任务变化检测，任务知识的记忆/检索，分类器的重组和知识转移。知识转移是通过基于任务相关性将所有源任务的度量转移到目标任务来实现的。实验结果表明，在所提出的RAN-MTLML中引入度量学习和度量上的知识转移是有效的。
In this paper, we propose a new sequential multitask pattern recognition model called Resource Allocating Network for Multi-Task Learning with Metric Learning (RAN-MTLML). RAN-MTLML has the following five functions: one-pass incremental learning, task-change detection, memory/retrieval of task knowledge, reorganization of classifier, and knowledge transfer. The knowledge transfer is actualized by transferring the metrics of all source tasks to a target task based on the task relatedness. Experimental results demonstrate the effectiveness of introducing the metric learning and the knowledge transfer on metric in the proposed RAN-MTLML.