论文标题
部分可观测时空混沌系统的无模型预测
Deep congruences + the Brauer-Nesbitt theorem
论文作者
论文摘要
储层计算是预测湍流的有力工具，其简单的架构具有处理大型系统的计算效率。然而，其实现通常需要完整的状态向量测量和系统非线性知识。我们使用非线性投影函数将系统测量扩展到高维空间，然后将其输入到储层中以获得预测。我们展示了这种储层计算网络在时空混沌系统上的应用，该系统模拟了湍流的若干特征。我们表明，使用径向基函数作为非线性投影器，即使只有部分观测并且不知道控制方程，也能稳健地捕捉复杂的系统非线性。最后，我们表明，当测量稀疏、不完整且带有噪声，甚至控制方程变得不准确时，我们的网络仍然可以产生相当准确的预测，从而为实际湍流系统的无模型预测铺平了道路。
We prove that mod-$p$ congruences between polynomials in $\mathbb{Z}_p[X]$ are equivalent to deeper $p$-power congruences between power-sum functions of their roots. This result generalizes to torsion-free $\mathbb{Z}_{(p)}$-algebras modulo divided-power ideals. Our approach is combinatorial: we introduce a $p$-equivalence relation on partitions, and use it to prove that certain linear combinations of power-sum functions are $p$-integral. We also include a second proof, short and algebraic, suggested by an anonymous referee. As a corollary we obtain a refinement of the Brauer-Nesbitt theorem for a single linear operator, motivated by the study of Hecke modules of mod-$p$ modular forms.
