学习与复合高清模型Learning to Learn with Compound HD Models |
|
课程网址: | http://videolectures.net/nips2011_salakhutdinov_hdmodels/ |
主讲教师: | Russ R Salakhutdinov |
开课单位: | 多伦多大学 |
开课时间: | 2012-09-06 |
课程语种: | 英语 |
中文简介: | 我们介绍了hd(或";hierarchical deep";)模型,这是一种新的复合学习体系结构,它将深度学习模型与结构化分层贝叶斯模型集成在一起。具体来说,我们展示了如何在深层Boltzmann机器(DBM)中的顶层功能活动之前学习分层Dirichlet过程(HDP)。这种复合HDP-DBM模型从很少的训练示例中学习新概念,通过学习低级通用特征、捕获低级特征之间相关性的高级特征,以及用于在不同类型概念的典型高级特征上共享优先级的类别层次结构。我们提出了hdp-dbm模型的有效学习和推理算法,并证明它能够从cifar-100对象识别、手写字符识别和人体运动捕捉数据集的很少实例中学习新概念。 |
课程简介: | We introduce HD (or "Hierarchical-Deep") models, a new compositional learning architecture that integrates deep learning models with structured hierarchical Bayesian models. Specifically we show how we can learn a hierarchical Dirichlet process (HDP) prior over the activities of the top-level features in a Deep Boltzmann Machine (DBM). This compound HDP-DBM model learns to learn novel concepts from very few training examples, by learning low-level generic features, high-level features that capture correlations among low-level features, and a category hierarchy for sharing priors over the high-level features that are typical of different kinds of concepts. We present efficient learning and inference algorithms for the HDP-DBM model and show that it is able to learn new concepts from very few examples on CIFAR-100 object recognition, handwritten character recognition, and human motion capture datasets. |
关 键 词: | 神经网络; 计算机科学; 机器学习; 贝叶斯学习 |
课程来源: | 视频讲座网 |
最后编审: | 2020-06-02:毛岱琦(课程编辑志愿者) |
阅读次数: | 42 |