图 书 馆
l1无穷正则化的一种有效投影An Efficient Projection for L1 Infinity Regularization |
|
| 课程网址: | http://videolectures.net/icml09_quattoni_epre/ |
| 主讲教师: | Ariadna Quattoni |
| 开课单位: | 加泰罗尼亚政治大学 |
| 开课时间: | 2009-08-26 |
| 课程语种: | 英语 |
| 中文简介: |  |
| 课程简介: | In recent years the L1,Infinity norm has been proposed for joint regularization. In essence, this type of regularization aims at extending the L1 framework for learning sparse models to a setting where the goal is to learn a set of jointly sparse models. In this paper we derive a simple and effective projected gradient method for optimization of L1,Infinity regularized problems. The main challenge in developing such a method resides on being able to compute efficient projections to the L1,Infinity ball. We present an algorithm that works in O(n log n) time and O(n) memory where n is the number of parameters. We test our algorithm in a multi-task image annotation problem. Our results show that L1,Infinity leads to better performance than both L2 and L1 regularization and that it is is effective in discovering jointly sparse solutions. |
| 关 键 词: | 联合正则化; 学习稀疏模型; 投影梯度 |
| 课程来源: | 视频讲座网 |
| 最后编审: | 2019-04-24：cwx |
| 阅读次数: | 49 |