(This piece was originally published as a guest post for The Quantum Insider)
The focus on quantum computing has been growing steadily over the last decade, thanks to the promise of how this technology could revolutionise our lives. An example of where quantum computing could be particularly groundbreaking is in the pharmaceutical industry, where it has the potential to make R&D drug development processes possible, that are currently impossible, due to the limitations of classical computing. This would enable vast reductions in the required resources to research and bring to the market new drugs, at a time when costs within the industry are majorly hindering the development of badly-needed new antibiotics and vaccines.
This promise is why quantum technology is already a driving force behind many innovation teams, with large companies such as IBM and Google dedicating specialist divisions to its development. These companies are already seeing results, with Google claiming in 2019 that its 54-qubit ‘Sycamore’ processor had performed a particularly complex calculation 158 million times faster than the world’s current, most powerful classical computer. Although some experts estimate that we are a decade away from a real breakthrough in quantum computing, there are others, like Goldman Sachs, who have predicted this could happen by 2026. In any case, it is now clear that the question for quantum computing technology is no longer “if”, but “when”, and once quantum computers become established, the increase in processing power will be exponential - with quantum computers quadrupling in power every 18 months.
Still some way to go
Despite the excitement around quantum’s numerous potential applications, we are still far from the stage where the technology can be easily applied to sectors across the board. In fact, the current state of quantum computing development is akin to traditional computers in the 50s and 60s. However, once the technology is made available, companies across different industries will need to develop excellence in how to run and make use of it for their particular product or service, if they are to maximally exploit its benefits.
Quantum computing is still in its early stages and is a complex technology to achieve. It is heavily resource intensive in many ways including the infrastructure and energy required to run both development and productionised systems, and the necessary talent to build and maintain them. There is an explosion of smaller companies trying to innovate and solve these challenges, but individual businesses exploring quantum computing ‘on their own’ are in many respects slower than larger corporations due to limitations in their knowledge, resources, and talent. This is the reason why collaboration is so essential for the quantum industry to progress and realise its full potential.
Collaboration is key
There are encouraging signs that some companies are taking steps to work more collaboratively. For example, in the financial sector the banking giant HSBC joined forces with 12 other companies through the NEASQC project to develop possible use cases for the technology, from carbon capture to energy infrastructure risk assessment, and breast cancer detection.
Despite this, expertise on quantum computing is still spread thinly across the country with little cooperation and coordination between the different, highly knowledgeable teams. Bringing this expertise together for the united growth of the industry is the key to making quantum computing a valuable reality. Governments can also play an important role in supporting this innovation, either with funds or scholarships to advance research into the industry, or by directly establishing centres of excellence.
Mind Foundry has had the privilege to enter into one such collaboration on a 3 year, £6.8 million InnovateUK funded project, AutoQT, led by Riverlane and with five other distinguished partners from the quantum sector: SeeQC, OxIonics, the National Physical Laboratory, the University of Edinburgh, and the University of Oxford. Together, these seven partners are solving one of the fundamental barriers that stands in the way of scaling the technology, to create devices with useful numbers of qubits: quantum device calibration.
Quantum devices are sensitive instruments. Apart from the difficulties involved in reliably manipulating a quantum mechanical system, that is sometimes as small as a subatomic particle, the devices themselves exhibit manufacturing variability within a single design, and the effect of environmental conditions is impossible to fully eradicate. The problem is far worse than the occasional cosmic ray flicking a transistor from a 1 to a 0 in a classical computer: it adds years to hardware development cycles, and as the problem grows with larger numbers of qubits on the same device, it also inhibits the production and operation of usefully scaled quantum computers.
Mind Foundry, together with these six partners, aims to solve this problem using Machine Learning by automating the calibration of quantum computing devices, and thus enabling a sovereign UK quantum industry to flourish.
Quantum computing could be a game-changing technology across multiple industries. Drug discovery could take hours instead of years. Highly complex models for fraud detection and risk assessment could happen in seconds. Simulations to model climate change that were unattainable with classical computing could finally become possible. But before we get there, everyone in this industry needs to start working together, aligning their financial and human resources in an effort to fully realise the potential of quantum computing. Experts estimate we are 10 years away from a true breakthrough in quantum computing, but if we work together we could be much closer.
If you'd like to read more about the work that is going into quantum research using AI, then check out Nathan's article on AI for Rapid Automated Calibration of Quantum Devices.
Want something more technical? Check out our case study describing how we created the world’s first hybrid quantum computing algorithm for training a generative model. The results of this work were also published in ScienceAdvances.