And anyone who makes such claims is hawking nothing but hype. Let us find out from a Maths wizard why things stand that way despite all the drum-roll around quantum computing. And why IBM does not subscribe to the idea of quantum supremacy. Pratima H explores.
Let’s think of the word ‘advantage’ before we go on tangents like ‘supremacy’ and ‘volume’ when we discuss adjectives around quantum computers. Let us also take into account facets like device crosstalk, the cloud context, entanglement and application muscle here. Asking Dr. Bob Sutor, Vice President, IBM Quantum Ecosystem Development about the on-the-street reality of quantum computers is like standing next to the box with the Schrodinger’s cat. Because after the third question one does something impossible—one forgets about the cat and starts to think of the box as a shoe-box. It’s not about the shoes. It never was. It’s what you do with them. Is that what he means? Dance around and find your feet.
Dr. Sutor, who has spent over two decades in IBM Research in New York, is also the author of Dancing with Qubits, a book about quantum computing. He has worked on, and led efforts, in symbolic mathematical computation, optimization, AI, blockchain, and quantum computing with his special hat on.
A theoretical mathematician by training, Sutor started coding when he was 15 and has got a hands-on métier in everything that stirred the software world from mathematics and mathematical software, Linux, open source, computer algebra, web standards to the current rockstars like quantum computing, AI, blockchain. How about some corridor walk with him chatting on what’s coming our way next?
If you were to explain the concept of a Quantum Computer from your lens, how would you describe it?
I start with the observation that all computing systems—quantum computers along with today’s classical computers such as phones, laptops, and supercomputers—operate on the ability to store and manipulate information. Classical computers manipulate individual bits, which store information as 0s and 1s. These bits can represent numbers, text, arbitrary data, and even what to draw on a screen.
Quantum computers, alternatively, use the physical phenomena of nature’s quantum mechanics to represent and manipulate information. Here we have quantum bits, or qubits. Unlike a bit, which has to be a 0 or a 1, a qubit can hold two pieces of information and so, in some sense, has two dimensions in which to operate. This is called superposition. The magic happens when we start using more qubits. For every extra qubit and using a fascinating property called entanglement, we double the information available for use. For example, just 50 qubits could represent more than one quadrillion compute states, simultaneously.
It’s important to understand that quantum computers are not a replacement for classical ones. They complement today’s computers with the potential to soon solve certain intractable problems that become extremely large or time-consuming during computation. Problems, for example, found in molecular simulation, or exponentially large optimization simulations in finance.
Can you share more details on the significance of circuits and cloud in accelerating Quantum Computing (QC)'s progress?
IBM put the first quantum computing device, open to the public, on the cloud in 2016. IBM was also the first to offer cloud-based commercial universal quantum computing systems as part of the IBM Q Network initiative, beginning in 2017. Today, we have 18 quantum systems available on the IBM Cloud. Our 240,000+ registered users, including those in the 100+ IBM Q Network organizations have run hundreds of billions of circuits on these systems, leading to more than 200 published research papers, and numerous industry case studies—all with the goal of developing practical quantum applications.
Quantum computing on the cloud is not new. We are now in our fifth year of providing it with the largest ecosystem, and development based on the community-driven open source Qiskit platform.
A circuit represents the basic unit of work for a quantum computer. Over four days in early May, our users ran over 1 billion circuits each day on actual, authentic quantum hardware.
Our team recently introduced circuit libraries into Qiskit to provide families of circuits for several practical applications and quantum algorithms. For example, the new optimization module is a library for researchers and beginners alike to experiment in quantum combinatorial optimization.
How does this compare to the 'number of qubits', 'superposition', 'entanglement' and other aspects that often come up as claims of a good quantum computer?
For clarity, qubits’ properties of superposition and entanglement, as well as interference, are what give them their quantum mechanical processing power.
The number of qubits is just one of many aspects that help determine the performance of a quantum computer. It’s why IBM developed the quantum architecture-agnostic metric, Quantum Volume.
Quantum Volume takes into account the number of qubits, connectivity, and gate and measurement errors. Improvements to underlying physical hardware, reduction of device crosstalk, and software circuit compiler efficiency can drive measurable progress in Quantum Volume.
IBM’s stated goal is to annually double Quantum Volume, which we have done since 2016. Our latest 32-Quantum Volume system, Paris, is a 27-qubit device available to organizations in the IBM Q Network.
What would determine real quantum supremacy-scale, speed, level of noise, fidelity or real-world problem-solving? Should we compare these computers to supercomputers or legacy computers or something else in measuring these metrics?
IBM does not subscribe to the idea of quantum supremacy. You can read our official statement on this. But what I can say here is that we are focused on Quantum Advantage: the research and development behind practical quantum applications that provide benefits to science and business that reach beyond what classical computers will ever be capable of alone. Quantum Volume, the metric described in my previous response, is a guidepost for the performance of quantum computers.
When can we see solid QC applications in the real world—how much has been done, could be done, with extreme projects like the Covid vaccine?
Quantum computers today cannot help develop a vaccine for COVID-19 since they are simply not powerful enough yet. Let me make a stronger statement: if you see claims of quantum computing helping COVID-19 research, those claims are pure marketing hype.
IBM has focused on bringing the best classical supercomputers to researchers and scientists helping to fight this virus. In the future, and I’m talking about at least 10 years from now, we expect this could be a problem for which quantum computing might be useful. More generally, simulation of some physical processes such as molecular reactions that help us create new materials may be available in the second half of this decade.
So what then are the constraints that still come in the way of QC applications?
To accomplish the development of applications with a quantum advantage, our scientists have collaborated with and published several research papers and use cases with IBM Q Network member organizations. This research includes several quantum finance-related research papers with partners including Barclays, JPMorgan, Wells Fargo, and MUFG Bank in Japan.
Much of the research we’ve conducted examines applications in risk analysis and option pricing. For example, in 2019, we published with JPMorgan the paper ‘Option Pricing Using Quantum Computers’ where we “presented a methodology to price options and portfolios of options on a gate-based quantum computer using amplitude estimation, an algorithm which provides a quadratic speedup compared to classical Monte Carlo methods.” And in 2019, with Barclays, we published ‘Quantum Algorithms for Mixed Binary Optimization applied to Transaction Settlement’ which introduced a novel approach for extending existing quantum methods for combinatorial optimization. The research optimizes the efficiency of securities settlements in capital markets, which is a difficult optimization challenge of importance due to the volume and value of transactions settled. Better algorithms could increase settlement efficiency, thereby minimizing the time period between trade and settlement.
In your reckoning, what would be the tipping point of QC, or have we already touched it? How does IBM feel distinct when others like Google, Microsoft, Baidu etc, are making big announcements every now and then claiming an edge of sorts?
The industry does not yet have a quantum computer capable of delivering a practical advantage over what classical computers can currently accomplish. We believe that an application with a quantum advantage will be developed within the decade for two main reasons:
First—IBM Quantum technology spans quantum hardware and software—all available on the cloud. IBM is the only company offering this complete stack.
Second—IBM Quantum’s global ecosystem includes more than 240,000 users experimenting with our public devices, more than 100 organizations with access to our premium devices, and academic collaborators helping educate and develop the needed skills for a larger, more diverse “quantum ready” workforce.
By any metric, IBM Quantum has the broadest ecosystem; the most advanced technology, and has made more advances than any other organization in the world.
Are issues like overheating, refrigeration, superconducting levels, propensity for errors, fragility, hardware availability etc. big challenges for the evolution of QC? What do you count as the big factor that needs to be addressed in the race of a great quantum computer?
Reducing noise and device errors are aspects of ongoing work that will continue for many years to develop applications with a Quantum Advantage. This will lead us to building fully fault-tolerant quantum computers. But accessing, experimenting, and developing near-term use cases are already possible on real, authentic quantum hardware. IBM offers 18 quantum computing systems over the cloud, many of which are publicly available for anyone to use, via the ‘IBM Quantum Experience’ platform.
Can there be something of a Moore’s law here as well?
We measure these systems’ performance by their Quantum Volume, explained in a previous answer. Quantum Volume is doubling every year in a similar Moore’s Law trajectory.
What else keeps you excited about QC? Any current work or experiments that you want to talk about?
I’m excited about how our community has grown. Just last month, on the fourth anniversary of IBM putting the first quantum computer on the cloud, our team hosted the virtual IBM Quantum Challenge: a series of exercises designed to educate people to use quantum computing in a fun way.
Over four days, 1,745 people from 45 countries came together and ran more than 1 billion quantum circuits per day on 18 quantum computers on the cloud to solve four problems ranging from introductory topics in quantum computing, to understanding how to mitigate noise in a real system, to learning about historic work in quantum cryptography, to seeing how close they could come to the best optimization result for a quantum circuit. This was record-breaking and historic event for quantum education and skills development.
The Quantum Computing Era on the Cloud is now in its fifth year. I’m excited to be part of the team at IBM Research which is making the most advanced systems available to the most users.