Abstract Quantum computers require precise control over parameters and careful engineering of the underlying physical system. In contrast, neural networks have evolved to tolerate imprecision and inhomogeneity. Here, using a reservoir computing architecture we show how a random network of quantum nodes can be used as a robust hardware for quantum computing. It induces quantum operations by optimising only a single layer of quantum nodes, a key advantage over the traditional neural networks where many layers of neurons have to be optimised. We demonstrate how a single network can induce different quantum gates, including a universal gate set. Moreover, we show that sequences of multiple quantum gates in quantum circuits can be compressed with a single operation, greatly reducing the operation time and complexity. As the key resource is a random network of nodes, with no specific topology or structure, this architecture is a hardware friendly alternative paradigm for quantum computation.
Quantum computers promise vastly enhanced computational power and an uncanny ability to solve classically intractable problems. However, few proposals exist for robust, solid-state implementation of such computers where the quantum gates are sufficiently miniaturized to have nanometer-scale dimensions. Here I present a new approach whereby a complete computer with nanoscale gates might be self-assembled using chemical synthesis. Specifically, I demonstrate how to self-assemble the fundamental unit of this quantum computer---a two-qubit universal quantum gate---based on two exchange coupled multilayered quantum dots. Then I show how these gates can be wired using thiolated conjugated molecules as electrical connectors. Each quantum dot in this architecture consists of ferromagnet-semiconductor-ferromagnet layers. The ground state in the semiconductor layer is spin split because of the Rashba interaction and the spin-splitting energy can be varied by an external electrostatic potential applied to the dot. A spin polarized electron is injected into each dot from one of the ferromagnetic layers and trapped by Coulomb blockade. Its spin orientation encodes a qubit. Arbitrary qubit rotations are effected by bringing the spin-splitting energy in a target quantum dot in resonance with a global ac magnetic field by applying a potential pulse of appropriate amplitude and duration to the dot. The controlled dynamics of the universal two-qubit rotation operation can be realized by exploiting the exchange coupling with the nearest-neighboring dot. The qubit (spin orientation) is read via the current induced between the ferromagnetic layers under an applied potential. The ferromagnetic layers act as ``polarizers'' and ``analyzers'' for spin injection and detection. A complete prescription for initialization of the computer and data input/output operations is presented. This paradigm, to the best of our knowledge, draws together two great recent scientific advances: one in materials science (nanoscale self-assembly) and the other in information science (quantum computing).
To implement a scalable and fault-tolerant programmable quantum computing architecture, the realization of high-fidelity, robust, and efficient quantum entangling gates in a multiqubit system is an essential requirement. In this paper, we consider the optimization of the segmented amplitude-modulated entangling gates realized by applying spin-dependent forces to a chain of ions trapped in a hybrid quadratic and quartic potential. We propose a geometrical hypersurface-tangent method to optimize the laser power and the pulse gradient, carefully considering the experimental feasibility in realizing the effective Rabi frequency and its change between adjacent pulse segments. In addition, our method allows us to optimize the solution for the case of a few segments by constructing an approximated null space through trading a negligible amount of gate fidelity. Finally, we show that the present scheme can provide a unified framework to improve the robustness to an arbitrary order against random static drifts of the motional frequencies, the gate duration errors, the laser detuning drifts, and their simultaneous drifts.