 |
Center for Nonlinear Studies |
Simulation protocols that hybridise tensor networks and stabilizer tableaus have garnered significant interest in recent years as they enable the simulation of circuits hitherto inaccessible to classical simulators as they are both highly entangled and contain extensive amounts of non-Clifford operations. By introducing useful resource states we have expanded the regime that hybrid Stabilizer Tensor Network (STN) simulators are capable of simulating. We demonstrate this by considering the paradigmatic benchmarking circuit of random T-doped N-qubit Clifford circuits, finding that we are able to simulate circuits with up to N T gates in polynomial time which is a significant improvement over the standard STN method as well as conventional MPS and stabilizer tableau simulators. Additionally, we demonstrate the specific conditions under which simulation is expensive using this protocol, which provides a heuristic for determining whether a given circuit is simulable. By way of application, we consider Quantum Error Correction (QEC) simulation. QEC circuits are usually simulated using stabilizer methods, as they consist of only non-Clifford gates. In contrast, the large amount of entanglement present in QEC circuits prevents simulation by tensor network methods. However, the use of stabilizer methods limits our ability to study QEC protocols using non-Clifford resources, such as logical T operations, or non-Clifford noise models. Using STN, we demonstrate fast simulation of QEC circuits with non-Clifford resources, which are intractable using traditional simulation techniques.
Bios: Azar Nakhl, Ben Harper and Max West are PhD students at the University of Melbourne, Australia.
Azar specializes in tensor networks, in particular hybrid models that incorporate techniques from stabilizer simulations.
Ben's interests include classical stabilizer simulation of quantum computation and the resource costs of quantum algorithms, as well as quantum error correction.
Max's research focuses on quantum computation and quantum information, specially on how to leverage tools from representation theory and random matrix theory to help explain both quantum systems and the algorithms that characterise them.
Host: Martin Larocca, T-4/CNLS