论文标题
部分可观测时空混沌系统的无模型预测
Robust cryogenic matched low-pass coaxial filters for quantum computing applications
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。
Electromagnetic noise is one of the key external factors decreasing superconducting qubits coherence. Matched coaxial filters can prevent microwave and IR photons negative influence on superconducting quantum circuits. Here, we report on design and fabrication route of matched low-pass coaxial filters for noise-sensitive measurements at milliKelvin temperatures. A robust transmission coefficient with designed linear absorption (-1dB/GHz) and ultralow reflection losses less than -20 dB up to 20 GHz is achieved. We present a mathematical model for evaluating and predicting filters transmission parameters depending on their dimensions. It is experimentally approved on two filters prototypes different lengths with compound of Cu powder and Stycast commercial resin demonstrating excellent matching. The presented design and assembly route are universal for various compounds and provide high repeatability of geometrical and microwave characteristics. Finally, we demonstrate three filters with almost equal reflection and transmission characteristics in the range from 0 to 20 GHz, which is quite useful to control multiple channel superconducting quantum circuits.
