ABSTRACT We model the hardware and software architecture for generalized Internet of Things (IoT) by quantum cloud-computing and blockchain. To reduce the measurement error and increase the efficiency of quantum entanglement (i.e. the capability of fault tolerance) in the current quantum computers and communications, we design a quantum-computing chip by modelling it as a multi-input multi-output (MIMO) quantum channel and obtain its channel capacity via our recently derived mutual information formula. To capture the internal qubit data flow dynamics of the channel, we model it via a deep convolutional neural network (DCNN) with generalized stochastic pooling in terms of resource-competition among different quantum eigenmodes or users. The pooling is corresponding to a resource allocation policy with two levels of competitions as in cognitive radio: the first one is on users’ selection in a ‘win–lose’ manner; the second one is on resourcesharing among selected users in a ‘win–win’ manner. To wit, our scheduling policy is the one by mixing a saddle point to a zero-sum game problem and a Pareto optimal Nash equilibrium point to a nonzero- sum game problem. The effectiveness of our policy is proved by diffusion modelling with theory and numerical examples.

中文翻译：

---

基于人工智能和区块链的量子计算：建模、容错和容量调度

摘要 我们通过量子云计算和区块链对广义物联网 (IoT) 的硬件和软件架构进行建模。为了减少当前量子计算机和通信中量子纠缠的测量误差和提高效率（即容错能力），我们设计了一种量子计算芯片，将其建模为多输入多输出（MIMO）量子信道并通过我们最近推导出的互信息公式获得其信道容量。为了捕获通道的内部量子位数据流动态，我们通过深度卷积神经网络 (DCNN) 对其进行建模，该网络具有广义随机池，以处理不同量子本征模式或用户之间的资源竞争。池化对应于一个资源分配策略，在认知无线电中具有两个级别的竞争：第一个是用户以“双赢”方式进行选择；第二个是在选定的用户之间以“双赢”的方式共享资源。也就是说，我们的调度策略是将鞍点与零和博弈问题和帕累托最优纳什均衡点与非零和博弈问题混合在一起。我们的政策的有效性通过扩散模型与理论和数值例子来证明。