稳定性兼容核:超越高斯性用于脉冲响应估计
Stability-compatible kernels: beyond Gaussianity for impulse response estimation
Gianluigi Pillonetto · Mauro Bisiacco
中文摘要
基于核的正则化用于线性系统辨识,使用半正定核既定义再生核希尔伯特空间(RKHS),也作为随机脉冲响应的协方差模型。RKHS 稳定性已有精确的核条件,而高斯先验下的几乎必然 BIBO 稳定性需要更严格的协方差条件。这些限制在高斯性之外未必持续存在,此时协方差不再决定完整概率律。引入稳定性兼容核(stability-compatible kernel)的概念,即存在至少一个几乎必然绝对可和样本路径的随机实现的协方差核。证明稳定性兼容性成立当且仅当稳定元素在关联 RKHS 中稠密,该条件严格弱于 RKHS 稳定性,并推导实用充分条件,包括平稳核的谱准则。进而构造非高斯先验,在保持指定协方差的同时强制几乎必然稳定性。特别地,这产生基于平稳径向基函数(RBF)核的稳定先验,将协方差结构与强制稳定性的机制分离。数值实验显示,对几类重要脉冲响应有显著估计增益,当相关动态出现在滞后晚期或分离时间区域时优势尤为明显。
关键要点
- 01指出高斯先验下几乎必然 BIBO 稳定性要求比 RKHS 稳定性更严格的协方差条件
- 02提出稳定性兼容核概念,等价于稳定元素在 RKHS 中稠密,条件严格弱于 RKHS 稳定性
- 03给出实用充分条件,包括平稳核的谱准则
- 04构造非高斯先验保持指定协方差并强制几乎必然稳定性,如基于平稳 RBF 核的稳定先验
- 05数值实验显示对滞后晚期或分离时间区域的脉冲响应估计增益显著
解读
尚无解读。
原始英文摘要
arXiv:2611.24272v1 Announce Type: new Abstract: Kernel-based regularization for linear system identification uses positive semidefinite kernels both to define reproducing kernel Hilbert spaces (RKHSs) and as covariance models for stochastic impulse responses. Sharp kernel conditions are known for RKHS stability, while almost-sure BIBO stability under Gaussian priors requires an even more restrictive covariance condition. These restrictions, however, need not persist beyond Gaussianity, where the covariance no longer determines the full probability law. We introduce the notion of a stability-compatible kernel, namely a covariance kernel admitting at least one stochastic realization with almost-surely absolutely summable sample paths. We show that stability-compatibility holds if and only if the stable elements are dense in the associated RKHS, a condition strictly weaker than RKHS stability, and derive practical sufficient conditions, including a spectral criterion for stationary kernels. We then construct non-Gaussian priors that preserve a prescribed covariance while enforcing almost-sure stability. In particular, this yields a stable prior based on the stationary radial-basis-function (RBF) kernel, separating covariance structure from the mechanism enforcing stability. Numerical experiments show substantial estimation gains for several important classes of impulse responses, with particularly pronounced advantages when relevant dynamics occur at late lags or in separated temporal regions.