摘要
To visualize structure separation for multiple tasks in neural networks, a learning strategy is proposed by incorporating the concept of neural pruning into the learning process. Visualization by placing the neurons in a 2D plane makes clear the task related structure in composite tasks. Experiment on a composite task using two tasks from UCI dataset show that the neural network of the composite task is composed of more than 80% neurons in each task. The proposed learning strategy aims the transfer learning of robot motion, and a partial experimental result is also referred.
源语言 | 英语 |
---|---|
出版状态 | 已出版 - 2012 |
已对外发布 | 是 |
活动 | 5th International Symposium on Computational Intelligence and Industrial Applications, ISCIIA 2012 - Sapporo, 日本 期限: 20 8月 2012 → 23 8月 2012 |
会议
会议 | 5th International Symposium on Computational Intelligence and Industrial Applications, ISCIIA 2012 |
---|---|
国家/地区 | 日本 |
市 | Sapporo |
时期 | 20/08/12 → 23/08/12 |
指纹
探究 'A pruning-based learning strategy to visualize self structureseparationin neural networks' 的科研主题。它们共同构成独一无二的指纹。引用此
Shibata, A., Lu, J., Dong, F., & Hirota, K. (2012). A pruning-based learning strategy to visualize self structureseparationin neural networks. 论文发表于 5th International Symposium on Computational Intelligence and Industrial Applications, ISCIIA 2012, Sapporo, 日本.