CRF的专业化和潜在的可能性Majorization for CRFs and Latent Likelihoods |
|
课程网址: | http://videolectures.net/machine_choromanska_majorization/ |
主讲教师: | Anna Choromanska |
开课单位: | 哥伦比亚大学 |
开课时间: | 2013-06-14 |
课程语种: | 英语 |
中文简介: | 分区函数在概率建模中起关键作用,包括条件随机场,图形模型和最大似然估计。为了优化分区函数,本文介绍了二次变分上界。这种不等式有助于主要化方法:通过更简单的子问题的迭代解决方案来优化复杂的函数。即使分区函数涉及图形模型(具有小树宽度)或潜在似然设置,这样的界限仍然有效地计算。对于大规模问题,提供了绑定的低级版本,并且优于LBFGS以及一阶方法。显示了几个学习应用程序,并简化为快速和收敛的更新规则。实验结果显示出优于现有技术优化方法的优点。 |
课程简介: | The partition function plays a key role in probabilistic modeling including conditional random fields, graphical models, and maximum likelihood estimation. To optimize partition functions, this article introduces a quadratic variational upper bound. This inequality facilitates majorization methods: optimization of complicated functions through the iterative solution of simpler sub-problems. Such bounds remain efficient to compute even when the partition function involves a graphical model (with small tree-width) or in latent likelihood settings. For large-scale problems, low-rank versions of the bound are provided and outperform LBFGS as well as first-order methods. Several learning applications are shown and reduce to fast and convergent update rules. Experimental results show advantages over state-of-the-art optimization methods. |
关 键 词: | 分区函数; 概率建模; 随机场 |
课程来源: | 视频讲座网 |
最后编审: | 2019-05-15:cwx |
阅读次数: | 52 |