arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 1902.05068stat.MLcs.CVcs.LG
收藏

非高斯统计模型的广义变分推断的收敛性

On the Convergence of Extended Variational Inference for Non-Gaussian Statistical Models

Zhanyu Ma, Jalil Taghia, Jun Guo

摘要:

变分推断(VI)是贝叶斯估计中广泛使用的框架。对于大多数非高斯统计模型,要找到一个易于分析的解来估计参数的后验分布是不可行的。最近,一种改进的框架,即扩展的变分推断(EVI),已被引入并应用于通过采用下限近似的变分目标函数来获得解析易处理的解决方案。本文讨论并比较了EVI实现所需的两个条件,即弱条件和强条件。在实际应用中,无论是强条件还是弱条件,EVI的收敛性都依赖于下界逼近的选取。一般来说,两种近似策略,单下界(SLB)近似和多下界(MLB)近似,可以应用于执行下界近似。为了澄清SLB和MLB之间的差异,我们还将讨论上述两种近似的收敛性质。广泛的比较的基础上,一些现有的基于EVI的非高斯统计模型。通过理论分析,论证了弱条件与强条件的区别。定性和定量的实验结果显示SLB近似的优点。

英文摘要:

Variational inference (VI) is a widely used framework in Bayesian estimation. For most of the non-Gaussian statistical models, it is infeasible to find an analytically tractable solution to estimate the posterior distributions of the parameters. Recently, an improved framework, namely the extended variational inference (EVI), has been introduced and applied to derive analytically tractable solution by employing lower-bound approximation to the variational objective function. Two conditions required for EVI implementation, namely the weak condition and the strong condition, are discussed and compared in this paper. In practical implementation, the convergence of the EVI depends on the selection of the lower-bound approximation, no matter with the weak condition or the strong condition. In general, two approximation strategies, the single lower-bound (SLB) approximation and the multiple lower-bounds (MLB) approximation, can be applied to carry out the lower-bound approximation. To clarify the differences between the SLB and the MLB, we will also discuss the convergence properties of the aforementioned two approximations. Extensive comparisons are made based on some existing EVI-based non-Gaussian statistical models. Theoretical analysis are conducted to demonstrate the differences between the weak and the strong conditions. Qualitative and quantitative experimental results are presented to show the advantages of the SLB approximation.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑