Speaker: 

Caroline Tornow

Title: 

Improving Quantum Applications with Pulse-Level Compilation

Abstract: 

The performance of near-term quantum algorithms, like the Quantum Approximate Optimization Algorithm, on state-of-the-art quantum computers is still significantly limited by noise. These algorithms are typically represented by quantum circuits in which unitary gates process the information. On the hardware level, these gates are implemented with calibrated control pulses. We demonstrate a novel pulse-efficient circuit transpilation methodology with Qiskit on IBM Quantum Computers, which scales cross-resonance entangling pulses to reduce the total duration of the quantum circuit in comparison to a Controlled-NOT (CNOT)-based quantum circuit. This procedure therefore makes a better usage of the finite qubit coherence time. By leveraging Cartan’s decomposition of SU(4) gates, we realize arbitrary pulse-scaled two-qubit gates and benchmark our technique on IBM Quantum devices using quantum process tomography. For almost all implementations we observe a significant error reduction of the pulse-efficient quantum gates in comparison to the respective CNOT gate-based implementations. As a sample application of the pulse-efficient methodology we implement circuits of a depth-one Quantum Approximate Optimization Algorithm applied to the Maximum Cut optimization problem for a non-hardware native 11-qubit graph. Here, we find that the circuit pulse duration is decreased by up to 52% and the error is reduced by up to 38%. 

Speaker: 

Clemens Giuliani 

Title: 

Variational Simulation of Quantum Circuits with Entangled-Plaquette States

Abstract:

While it is largely believed that the classical simulation of general large quantum circuits is hard to achieve it is often the case that specific quantum circuits can be approximated with classical variational algorithms. Variational representations used so far comprise tensor networks as well as neural network quantum states based on shallow architectures. In this work we introduce a simulation strategy for quantum circuits based on a different Ansatz called entangled- plaquette states (EPS), which have previously been used for simulating quantum systems on a lattice. Within this representation, we outline which classes of quantum gates can be applied exactly or approximately and give examples for both qubit and photonic quantum circuits. As an application we demonstrate that EPS can in principle be used to simulate the quantum approximate optimization algorithm. Furthermore we extend the previous variational fidelity optimization to wavefunctions which can be zero and present an implementation of the stochastic reconfiguration optimization algorithm with automatic differentiation.