# Top 10 quantum computing experiments of 2019

The last decade has seen quantum computing grow from a niche research endeavour to a large-scale business operation. While it’s exciting that the field is experiencing a surge of private funding and media publicity, it’s worth remembering that nobody yet knows how to build a useful fault-tolerant quantum computer. The path ahead is not “just engineering”, and in the coming decade we have to pay attention to all the “alternative approaches”, “crazy ideas” and “new ways of doing things”.

With this in mind, I created this subjective list of quantum computing research highlights of 2019. It highlights **experimental achievements **which show new exciting ways of controlling qubits. In such a vast space of literature, I have no doubt I missed some essential works, so I encourage you to get in touch and add your favourites to the list.

Here are the highlights in no particular order.

## 1. Encoding logical qubits as grid states

P. Campagne-Ibarcq *et al, *A stabilized logical quantum bit encoded in grid states of a superconducting cavity,

C. Flühmann *et al, *Encoding a qubit in a trapped-ion mechanical oscillator, *Nature* vol. 566, pp. 513–517(2019).

For quantum computing to reach its full potential, it needs to become fault tolerant. However, the overhead full error correction is quite daunting, as thousands of physical qubits are used to encode one logical qubit — and even then, thousands of logical qubits need to be entangled to perform a useful computation. These works show a less scary path forward, where a logical qubit is encoded with just one harmonic oscillator, thus dramatically reducing necessary resources.

C. Flühmann and her colleagues at Jonathan Home’s group at ETH Zurich charted the path in early 2019 by manipulating such a qubit encoded into vibrational states of a trapped ion. Later in the year, work from P. Campagne-Ibarcq and colleagues at Yale applied this approach to create GKP (grid) states of microwave photons in a superconducting cavity. They managed to not only manipulate the logical qubit, but stabilise it permanently!

The superconducting qubit community has been debating for a while whether qubits should be encoded into transmons or as cavity photons (listen to this podcast to hear more), and it seems both approaches are going strong into the 2020.

## 2. Sneaky error mitigation by IBM

Abhinav Kandala *et al*, Error mitigation extends the computational reach of a noisy quantum processor*, Nature* vol. 567, pp. 491–495(2019)

Of all the work to come out of IBM quantum team in 2019, this one makes me smile the most. The idea seems simple yet brilliant. The researchers ask the question: given the outcomes of noisy quantum operations, can we predict what the outcomes would have been without the noise? The solution: increase the noise and extrapolate down to zero. And how do you increase noise? Make the gates the same way as always, simply slower.

## 3. 2D cluster states of photons

Mikkel V. Larsen *et al*, Deterministic generation of a two-dimensional cluster state, Science 18 Oct 2019: Vol. 366, Issue 6463, pp. 369–372.

Warit Asavanant *et al*, Generation of time-domain-multiplexed two-dimensional cluster state, Science 18 Oct 2019:

Vol. 366, Issue 6463, pp. 373–376.

Going against the popular tide of dual-rail photonic qubits, two groups came up with awesome ways of creating two-dimensional cluster states of photons entangled in continuous degrees of freedom. Cluster states are a resource for measurement-based quantum computation (MBQC), where arbitrary unitaries are constructed as sequences of measurements consuming the entanglement in the initial cluster. A reasonably simple scheme allowed Mikkel V. Larsen and colleagues to entangle over 30,000 photonic modes with each other, really bringing home the power of CV encodings.

## 4. Quantum s***remacy and its enemies

F. Arute *et al*, Quantum supremacy using a programmable superconducting processor, *Nature *Vol*. *574, pp. 505–510(2019) .

IBM’s take:

Of course this had to be here, right? Long anticipated result from a magnificent 53 qubit machine from the lab of John Martinis did not disappoint. It was a momentous event for the quantum computing community (and for science at large?), sparking a lot of discussion, both in- and outside academia.

Google’s state-of-the-art device applies random 1- and 2-qubit gates on 53 qubits. After 20 cycles, the fidelity drops below 1% — yet amazingly, this still gives a clear signal of quantum advantage. But those numbers also remind us how badly we need error correction — a computer which makes a mistake after few tens of operations is really not what most users have in mind…

Mysteriously, Google’s publication was immediately followed by IBM’s rebuttal — what a coincidence they had it ready! I believe most people would say Google’s result more or less stands — see Scott Aaronson’s blog for more discussion. However, I think IBM folks are right on point when they write: “*more fundamentally, […] quantum computers will never reign “supreme” over classical computers, but will rather work in concert with them, since each have their unique strengths”.*

Amen.

## 5. Boson sampling and its enemies

Hui Wang *et al*, Boson Sampling with 20 Input Photons and a 60-Mode Interferometer in a 10¹⁴-Dimensional Hilbert Space, Phys. Rev. Lett. 123, 250503 — Published 18 December 2019

Raúl García-Patrón *et al*, Simulating boson sampling in lossy architectures, Quantum 3, 169 (2019).

The runner-up to the supremacy result is this boson sampling work from Jian-Wei Pan’s group. They relentlessly pushed the efficiencies of single-photon sources over the years, and now constructed an interferometer which “almost” demonstrates quantum advantage: *“ theoretically calculating the full probability distribution in the 20-photon-input 14-output bosonsampling will take hours using supercomputers”. *And I mean, look at the optical table, it’s incredible this works!

But, while ideal boson sampling is conjectured to be a hard calculation, does the same hold for noisy boson sampling? A theoretical result published this year by Raúl García-Patrón and colleagues suggests the answer is “no”! The abstract is clear enough about this: *“In this work we show that using classical computers, one can efficiently simulate multi-photon interference in all architectures that suffer from an exponential decay of the transmission with the depth of the circuit, such as integrated photonic circuits or optical fibers”.*

So does it even make sense to continue building larger and larger boson samplers? I guess 2020 will bring more discussion of this issue.

## 6. Diamond qubits shine bright

C. E. Bradley *et al*, A Ten-Qubit Solid-State Spin Register with Quantum Memory up to One Minute, Phys. Rev. X 9, 031045 — Published 11 September 2019

Loophole-free Bell tests are an old story now, but one should not forget the power of NV centers just yet! Besides exciting applications in sensing and medicine, the research from Delft reminds there is much on offer for pure quantum information processing. The researchers generate arbitrary 2-qubit entangled states of a 10-qubit register — something that very few quantum computing labs in the world would find easy! Watch out for solid state systems in 2020.

## 7. Error correction with superconducting qubits

Christian Kraglund Andersen *et al*, Repeated Quantum Error Detection in a Surface Code, arXiv:1912.09410

Christian Kraglund Andersen *et al, *Entanglement stabilization using ancilla-based parity detection and real-time feedback in superconducting circuits, npj Quantum Inf 5, 69 (2019)

Despite all the progress in continuous variable encodings, transmon qubits in a surface code remain one of the leading architectures for error correction. And at the end of this year, C. K. Andersen and colleagues at ETH Zurich brought this goal an inch closer. They created a single square of a surface code — 4 physical qubits and 3 ancillas —, initialised a logical qubit and demonstrated measurements of all stabilisers. And in a separate experiment, the same group demonstrated the ability to do fast feedback conditioned on measurement outcomes.

These results follow previous experiments on trapped ions on the 7-qubit code from 2014 and stabiliser measurements with fast feedback from 2018. We did a podcast about the topic earlier in 2019.

## 8. Silicon qubits entangle, and they’re fast

Y.He *et al*, A two-qubit gate between phosphorus donor electrons in silicon. *Nature* Vol. 571, pp. 371–375 (2019)

Did I mention to watch out for solid state systems? Oh yes! Electrons in silicon are still in their infancy, but rapidly advancing computational capabilities. And 2019 marked the first demonstration of a two-qubit gate with donor atoms in silicon. And while the fidelity was a modest 90%, it only took 0.8 ns — p̵r̵o̵b̵a̵b̵l̵y̵ ̵t̵h̵e̵ ̵f̵a̵s̵t̵e̵s̵t̵ ̵t̵w̵o̵-̵q̵u̵b̵i̵t̵ ̵g̵a̵t̵e̵ ̵e̵v̵e̵r̵?̵ one of the fastest two-qubit gates ever.

These are just the beginnings, but since silicon is such a familiar material to the integrated circuit industry, the whole field can easily explode given the first knowledge.

## 9. Ion traps scale up with photonic networks

L. J. Stephenson *et al*, High-rate, high-fidelity entanglement of qubits across an elementary quantum network, arXiv:1911.10841

One of the main ideas for scaling up ion traps — from small processors to large-scale devices — is to connect individual units of few qubits with photonic links. The challenge is to make the connections fast and high-fidelity. The best previous demonstration achieved remote entanglement with fidelity of 78% and produced 4.5 entangled pairs per second.

This work from the researchers in Oxford pushes the numbers by some orders of magnitude! Their demonstration entangles ion in separate processors with fidelity of 94% and at an impressive rate of 182 pairs per second.

## 10. What is your favourite?

Quantum computing is a vibrant and diverse field, so no doubt this list misses some amazing experiments from 2019. What would you put as number 10? You can get in touch with me on Twitter @quantumpod or make a comment below.