Quantum Future | Part II: Long road ahead for quantum-optimised AI-machine learning

“AI can help as a tool for interpreting or reasoning some of the quantum processes that we see or the results that we’re getting, and in turn help develop quantum algorithms.” — Mitchell Stott

In part one of our series, “Quantum future”, we focused on the current foundational state of quantum computing, Australia’s significant role in its development, and the challenges and potential for IP protection in the space.

One popular question when discussing quantum computing relates to another rapidly advancing technology – artificial intelligence (AI). Due to the ability of quantum to compute millions of variables at once and efficiently digest highly complex data, innovators have pondered how this application could enhance AI-machine learning.

Put simply – we’re not there yet.

AI vs quantum – technological opposites

“AI is kind of the opposite of quantum computing as I see it,” says Dr Simone Shu-Yen Lee, a principal and patent attorney at Griffith Hack with a PhD in quantum optics. “Quantum computing is a computer. By using specific algorithms, we could run specific models to solve complex problems in fields like chemistry, physics, geological problems and drug discovery. They have the potential to more precisely model complex systems like weather, how our bodies react to drugs, pharmaceutical compositions, molecular interaction, traffic optimisation… the list goes on.”

“Quantum computing allows us to model things that have billions or trillions of variables or transactions,” adds Mitchell Stott, a patent attorney at Griffith Hack. “This just cannot be done on a classic computer like the ones on which AI currently operates. For example, weather simulations. Currently it’s very challenging to predict the weather in a week’s time because the number of variables make it impossible to make a highly accurate prediction. A quantum computer can run these variables simultaneously, or with less computation needed to reach the solution because of the fundamental aspect of qubit superposition, and the algorithms in place to take advantage of that for computation.”

“AI is different. Instead of doing exact computations, what it does is more like intuition and pattern recognition. AI takes a lot of data and considers what is more probable to happen based on that data set. The two technologies use two different kinds of thinking or computing. Marrying the two isn’t quite there yet,” Lee states.

AI as a support system

“I think there is potential for AI to assist in how we get to a quantum computer more quickly,” says Lee.

As discussed in part one of our series, innovation in quantum computing is in its infancy, and innovators are working towards the most optimal development of the qubits on which this technology relies. This involves the materials they are built on, navigating how to get them to function at or near room temperatures, reducing size, error reduction and noise amongst other functional areas. This is where experts see the highest potential for AI support.

“Control systems and error correction would likely be where the strength of AI lies, essentially improving the processes in trying to develop patterns to best enhance processes which will enable quantum computers to work more efficiently,” says Lee.

“AI can help as a tool for interpreting or reasoning some of the quantum processes that we see or the results that we’re getting, and in turn help develop quantum algorithms,” adds Mitchell.

Quantum and AI in practice

Innovators across quantum computing are recognising the potential of AI to further progress technology in the space, with startups and groups in Australia building strong connections.

“Recently, the Quantum Australia group has formed, supported by funding from the Australian government. Their aim is to band together different organisations both within the quantum technology space and other industries such as the health sciences – a strategic alignment due to the potential value of quantum innovation for health. These organisations are considering using AI in their processes to solve problems such as which drugs might be better used to address different health issues. This connection can help innovators consider the specific ways in which quantum computing can be applied, and then strategically develop that science,” says Lee.

Large tech companies such as US-based NVIDIA are also implementing AI to accelerate the roadmap to useful quantum computing, investigating its value for areas such as algorithm development and hardware control.

Just wait and see

The potential for coupling AI and quantum technologies is certainly there and could be very exciting – but before a scalable quantum computer is developed, it’s hard to predict what the outcome will be.

“This is an amazingly interesting and fast-moving field, and we can’t really predict what’s going to happen next,” says Lee.

“It’s a wait and see scenario. Once the first working scalable quantum computer is developed, things are going to happen very quickly. Then the nexus between AI and quantum computing will really reveal itself,” she concludes.