AWS unboxes quantum cat qubit kit called Ocelot

Sprinting after Microsoft and co, Amazon claims it too has a QC chip that's good at all-important error correction

Amazon Web Services on Thursday announced Ocelot, a quantum computing chip based on "cat qubits."

The feline moniker is a reference to Erwin Schrödinger's famous thought experiment involving a cat in a box with a device that has an equal chance of killing or not killing the animal at the end of an hour.

The point of that hypothetical scenario was to illustrate the indeterminacy of a quantum system – the cat would be both alive and dead to an external observer until the box was opened, in the same way that a quantum system exists in all possible states until it is measured.

The analogy also might be applied to quantum computing – it's both alive and dead at the same time.

Unless you're talking about quantum annealing, a specialized form of quantum computing commercialized by D-Wave, no one is really using generalized quantum computing technology for production workloads at the moment. But big names like Amazon, Google, IBM, and Microsoft believe that's a possibility – some day. It's the quantum cloud Field of Dreams: If you bill it, they will compute.

Microsoft last week showed off its Majorana 1 quantum computing chip, crowing that its topological qubits promise practical quantum computing in a matter of years rather than decades. There's still some skepticism about the IT giant's claims but for the sake of moving on, let's just assume there's something there.

Qubits are just quantum-mechanical systems (involving atomic particles) with a state, and thus come in various forms. Topological qubits are based on the topological properties (how elements are organized or connected) of materials, specifically Majorana particles. Photonic qubits are based on the quantum properties of light, like polarization and phase. AWS currently offers a quantum computing research platform called Braket that's based on trapped ion qubits.

Now, AWS has devised cat qubits which represent the state of boson (photon) oscillations – changes in amplitude and phase. Hence the name Ocelot.

Close-up of AWS's Ocelot quantum chip ... Source: Amazon. Click to enlarge

"Cat qubits use the quantum superposition of classical-like states of well-defined amplitude and phase to encode a qubit’s worth of information," the web goliath said.

The appeal of cat qubits – beyond the marketing advantage that comes from a cuddly term – is that they help with error correction.

"Increasing the number of photons in the oscillator can make the rate of the bit-flip errors exponentially small," Amazon explains. "This means that instead of increasing qubit count, we can simply increase the energy of an oscillator, making error correction far more efficient."

Error correction is the major stumbling block in quantum computing at the moment. Qubits are sensitive to all manner of potential disturbances – thermal, electromagnetic, seismic, etc – and require orders of magnitude more physical qubits – embodied in hardware – to correct the logical qubits used for programming. As a result, the space required to house the projected error correction capacity would be prohibitive – we're talking airplane hangars.

Ocelot, as Amazon claims in a Nature paper, handles error correction particularly well. It's a prototype quantum computing chip made of two integrated silicon microchips, each with an area about one square centimetre. The chips have a layer of superconducting material that form quantum circuit elements and consist of 14 core components. These include: Five data qubits (the cat qubits), five buffer circuits, and four more qubits for error detection. Ocelot's oscillators are made of tantalum, a superconducting material.

"Today Nature published measurement results from Ocelot, the new quantum chip created at the AWS Center for Quantum Computing at Caltech," John Preskill, professor of theoretical physics at the California Institute of Technology (Caltech) and an Amazon Scholar, told The Register. "There is still far to go, but we hope that Ocelet’s unique architecture will shorten the path to quantum utility that benefits the world."

There is still far to go, but we hope that Ocelet’s unique architecture will shorten the path to quantum utility that benefits the world

Heather West, researcher manager for quantum computing at IDC, said it's difficult to say whether AWS's Ocelot can help make quantum computing commercially meaningful in the next five years.

"With regards to Ocelot, I see it as an advancement," West told The Register. "We're at this time where we're starting to see this pivot from talking about the number of qubits to higher quality qubits and error correction.

"And through this pivot, we've gone first to using software to minimize some of the error – whether it be error mitigation, or error suppression – then we went to logical qubit discussions – and we're still there – but now we're talking about hardening the qubits themselves in order to reduce some of the error right off the bat."

West however said that a lot of unanswered questions remain.

"Amazon notes in their press release and in their articles that they see it as a scalable architecture, but they haven't really given any demonstration of how they plan to scale it," she explained.

"With a superconducting chip, you need to have wires for input and output for each qubit. As you scale up those wires, [the system] obviously scales exponentially as well. So that leads to size and footprint."

West also pointed to the high level of errors when superconducting is involved.

"Even though they're talking about the bit flip, you still have the phase flip that you need to worry about and the errors there. And then any other noise that comes into the system that you have to talk about. So talking about an error-corrected qubit is a good first step, but I think there might be a longer way to go. How long will that take to happen? I still think we're talking years away." ®