人工智能培训

搜索

深度学习论文:用于多输出高斯过程的非线性过程卷积(Non-linear process convolutions for multi-output Gaussi

[复制链接]
yyshrrrr 发表于 2018-10-11 09:02:08 | 显示全部楼层 |阅读模式
yyshrrrr 2018-10-11 09:02:08 318 0 显示全部楼层
深度学习论文:用于多输出高斯过程的非线性过程卷积(Non-linear process convolutions for multi-output Gaussian processes)本文介绍了用于构建多输出高斯过程的协方差函数的过程卷积形式的非线性版本。通过Volterra系列引入非线性,一系列pereach输出。我们在Volterra系列的输出处提供了近似高斯过程的均值函数和协方差函数的闭合表达式。联合高斯过程的均值函数和协方差函数是使用高斯变量的乘积矩的公式导出的。我们将非线性模型的性能与经典过程卷积方法在一个合成数据集和两个实时数据集中进行了比较。
The paper introduces a non-linear version of the process convolutionformalism for building covariance functions for multi-output Gaussianprocesses.The non-linearity is introduced via Volterra series, one series pereach output.We provide closed-form expressions for the mean function and thecovariance function of the approximated Gaussian process at the output of theVolterra series.The mean function and covariance function for the jointGaussian process are derived using formulae for the product moments of Gaussianvariables.We compare the performance of the non-linear model against theclassical process convolution approach in one synthetic dataset and two realdatasets.深度学习论文:用于多输出高斯过程的非线性过程卷积(Non-linear process convolutions for multi-output Gaussian processes) GGSCP7Z6SWHsdsmQ.jpg
URL地址:https://arxiv.org/abs/1810.04632     ----pdf下载地址:https://arxiv.org/pdf/1810.04632    ----深度学习论文:用于多输出高斯过程的非线性过程卷积(Non-linear process convolutions for multi-output Gaussian processes)
回复

使用道具 举报

您需要登录后才可以回帖 登录 | 立即注册

本版积分规则 返回列表 发新帖

yyshrrrr当前离线
新手上路

查看:318 | 回复:0

快速回复 返回顶部 返回列表