With a diverse range of speakers from both academia and industry, we were exposed to a wide range of topics across all three days of QCTIP2020. Most of the talks were recorded and can be found on Riverlane’s YouTube channel. Here we report on two key themes from the conference – quantum algorithms and error correction – and highlight findings from some of the talks.
Talks on quantum algorithms encompassed aspects of algorithm design, implementation and performance as well as potential improvements to existing algorithm paradigms. Many of the talks focused on current or NISQ-era algorithms. But some looked further into the future. Barbara Terhal (TUDelft) gave a keynote talk on quantum phase estimation applied to stoquastic Hamiltonians, and Birgitta Whaley (UC Berkeley) spoke about new methods to design quantum algorithms, such as using optimal quantum control to sequence quantum gates effectively.
Simon Benjamin (University of Oxford) gave an in-depth talk on the quantum computing emulator his group has developed, which includes efficient gradient descent algorithms and error mitigation. Kristan Temme (IBM Research) focused on how quantum algorithms can be used for many-body physics problems. Craig Gidney (GoogleQuantum AI) spoke on his method for solving 2048-bit RSA integers on noisy qubits — an area that quantum computation is expected to have significant impact on in the future. Mario Szegedy (Alibaba Cloud Quantum Lab) closed the conference with a fascinating talk on the quantum approximate optimisation algorithm (QAOA) and its limitations.
From our contributed speakers, which included doctoral students, postdocs and industry researchers, the talks covered topics such as efficient measurement in the presence of sampling errors (Ophelia Crawford, Riverlane), using the variational quantum eigen solver algorithm to solve theFermi-Hubbard model (Lana Mineh, University of Bristol), quantum circuit optimisation (John van de Wetering, Radboud University) and using the QAOA algorithm to tackle anon-planar graph problem, implemented on a superconducting processor (Matthew Harrigan, Google Quantum AI).
In terms of error correction, Aleksander Kubica (Perimeter Institute) presented some interesting joint work with Nicolas Delfosse. He described how to use any decoder for the2d toric code as a decoder for the 2d colour code. The principle idea is that by restricting one of the colours of the colour code, one recovers a toric code, and using two of these restrictions is sufficient for decoding the colour code.
Nicolas Delfosse (Microsoft) discussed some joint work with several co-authors that was of interest to both hardware enthusiasts and theorists. He described a micro-architecture for the Union-Find decoder implemented on the surface code considering a large-scale fault tolerant system. Taking inspiration from the assembly line in car manufacturing, Nicolas described a three-stage fully pipelined hardware implementation of the decoder that significantly speeds it up. By sharing resources between logical qubits, Nicolas considerably reduced the number of hardware units and memory capacity necessary for decoding. Using these improvements, he gave some numerical evidence that this micro-architecture can be executed fast enough to correct errors in a quantum computer.
Christopher Chamberland (AWS) delivered a talk on efficient flag decoding algorithms for topological codes. The motivation for this problem comes from architectures based on fixed frequency transmon qubits, where if two neighbouring qubits have the same frequencies they cannot interact. This suggests looking at topological codes with low nearest neighbour connectivity. Christopher demonstrated how flag qubits can be used to decode such codes.
Our final keynote speaker of the conference, Matthias Troyer (Microsoft Research), spoke on the trajectory of quantum computing and future obstacles to overcome, both in terms of hardware and software.
There was a huge breadth of work presented at QCTIP2020, giving an excellent overview of the current state-of-the-art in practical quantum computation. Do check out Riverlane’s YouTube channel to see QCTIP2020 for yourself.