Quantum Supremacy: are we nearly there yet?
Reflections on the Quantum for Business Conference in Silicon Valley
Syed Adil Rab, Quantum Computing Lead, Cogisen
Ok the headline of my blog is a little provocative, but it was a subject of much heated debate when I attended the Quantum for Business Conference in December. The event was a forum to bring together academics with venture capitalists, start-ups like Cogisen and some of the major Silicon Valley technology companies. The mission was to help identify practical, commercially viable applications for quantum computing and the discussion certainly did not disappoint. So, did the brain trust of the quantum computing community believe “quantum supremacy” is any closer?
Attendees from both the scientific community and investment companies argued that reaching this state will be the turning point for investment in quantum information science from major private and public sector institutions. And it is clear from presentations by IBM, Google and Microsoft showed that there is a clear determination to deliver a commercially viable quantum computer capable of operating at significant scale. In reality, though, none of the presentations moved beyond a theoretical demonstration of a quantum speed-up versus a classical algorithm and we are still some way off a fully-fledged quantum error corrected machine, which could operate such calculations.
There was some agreement that we are not that far away from seeing working applications of quantum computing. Cryptography was accepted by many as being the most likely first use case followed by simulation systems, which could understand many-body systems and electronic states of molecules – something that is of huge interest to bio-medical and pharmaceutical companies. Optimization using quantum resources to find optimal solutions in industries such as financial services was not viewed as possible using quantum computers, but there are quantum algorithms that are able to find much better solutions in exponentially less time. There was also some speculation on the use of quantum computers in machine learning, but attendees agreed that such systems would require large scale quantum computers with a large number of qubits. Everyone agreed that this was some way off and one of the key barriers to overcome was finding the right quantum hardware foundation for such applications.
It was symbolic of the stage we are at that there were more than 10 companies at the conference talking about 6 different ways to address the quantum hardware question. Superconducting was the preferred choice among a number of delegates, but there was also discussion of NMR, Majorana Fermions, FQH state, Trapped Ions and Quantum Dots. The main reason I was attending was to present some breakthrough research we have undertaken around the use of machine learning to improve the performance of large scale quantum dots. We are proposing a novel machine learning technique based on data remapping, which is not based on neural networks, but maps raw data into a space where information is linearised and can be extracted with dramatically less optimisation parameters. I will go into this in more detail in another blog, but we were very excited by the reaction we got at Quantum for Business.
Overall, the conference confirmed my belief that commercially viable quantum computing systems are achievable. It’s going to be really hard to get there and it is commonly agreed that will harder than the media hype last year suggested, but I am more confident than ever in my view that we can deliver on this goal. The biggest problems right now from an engineering point of view are scalability and noise. How can you scale up quantum hardware that has high fidelity (in simple terms that does what it is supposed to do) and emits low noise? And can it be scaled to thousands of qubits with relative ease? You might say that right now we don’t even have one single standard technology that holds the promise of becoming the foundation for a universal quantum computer – the debate at the conference about quantum hardware underlined this point. Soon enough, though, I am confident we will have a quantum machine capable of performing outstanding feats that give us incredibly interesting results. When companies like Airbus attend conferences like Quantum for Business and highlight their Quantum Computing Challenge it underlines how seriously big business is taking this technology. It will still be some time before it enters the mainstream properly, but we are beginning to see solutions to the challenges I’ve mentioned, which will enable manufacturers to build quantum machines with more qubits capable of processing larger calculations than classic computers. Quantum supremacy may not be so far away after all!
Syed Adil Rab is the Quantum Computing Lead at Cogisen. He holds a Doctorate in Quantum Information and Optics from Sapienza Università di Roma.
Cogisen Launches Machine Learning Accelerator.