IBM Watson CTO: Quantum computing could advance artificial intelligence by orders of magnitude

Combining the vast processing power of quantum computers with cognitive computing systems like IBM's Watson will lead to huge advances in artificial intelligence, according to a C-level executive at the US software giant.

Speaking to IBTimes UK at the recent Hello Tomorrow conference in Paris, IBM Watson's chief technology officer Rob High said there was a "very natural synergy" between cognitive computing and quantum computing, revealing he hoped to one day see Watson run on a quantum system.

What is quantum computing?

Quantum computers replace traditional bits that are used in digital communications with quantum bits, or qubits. Potential applications can be found in a variety of fields, from medicine to space travel.

Qubits exist in a state of superposition, meaning they can be in both states at once, rather than restricted to either binary state as traditional bits function.

"I'd love to see a quantum Watson," he said. "IBM Research is actually working on next generation computing. I can't say exact numbers but a [quantum Watson] would be orders of magnitude more powerful than systems that are currently being used.

"Besides the obvious that both cognitive computing and quantum computing depart substantially from the classical forms of information computing, the biggest synergies lie in the realisation that increasingly sophisticated reasoning strategies employed by cognitive computing are going to require increasingly powerful and efficient underlying computing architectures."

IBM's Watson supercomputer first rose to prominence in 2011 when it became the first computer to beat human contestants at the US gameshow Jeopardy!

In the years since, IBM and other companies have put Watson's immense computing power to a variety of uses, from working with doctors to develop treatment plans for cancer patients, to assisting the world's media in crunching tennis statistics at Wimbledon.

Quantum Watson?

IBM is yet to announce plans to integrate a quantum computer system with Watson but the software giant recently unveiled a new superconducting chip that demonstrates a technique crucial to the development of quantum computers.

The chip was a leap forward in research into quantum computers, as it was the first to integrate quantum bits – or qubits – into a two-dimensional grid. This is important for making a practical machine but there is still a long way to go before quantum computers find practical use.

Nasa, Google and the CIA are among the companies and organisations also working on quantum computers, while the UK government has outlined a £270m ($420m) strategy into quantum technology growth through the UK National Quantum Technology Programme.

Nasa's Quantum Artificial Intelligence Laboratory (QuAIL) is working specifically on assessing the potential quantum computers have with regards to artificial intelligence, though the agency is hazy on what exactly any machine might be used for beyond helping "address Nasa challenges".

Quantum computers have already been used to test artificial intelligence by researchers in China, albeit in a very limited capacity. Earlier in 2015, a team from the country's University of Science and Technology developed a quantum system capable of recognising handwritten characters in a demonstration they dubbed quantum artificial intelligence.

"Due to the widespreading importance of artificial intelligence and its tremendous consuming of computational resources, quantum speed up would be extremely attractive against the challenges from the Big Data," the abstract for the team's research paper stated.

This demonstration was on a quantum computer using only four qubits, leading to speculation of what a system using hundreds – or even thousands – of qubits would be capable of. Such machines do not yet exist, at least not commercially, but Canada-based quantum computing firm D-Wave systems recently claimed it has built a 1,000 qubit quantum computer.

According to Seth Lloyd, a professor of mechanical engineering at the Massachusetts Institute of Technology (MIT), a machine of just 300 qubits could be used to "map the whole universe", processing all the information that has existed since the Big Bang.

For High, introducing quantum computers to the field of cognitive computing machines such as IBM's Watson represents a logical step in realising the full benefits of both.

"Cognitive computing does emulate some aspects of human observation, interpretation and evaluation," he said. "While cognitive systems do not fully simulate how the human mind actually works, if they're to be useful in amplifying human cognition, cognitive systems will have to perform with increasing speed, agility and with lower energy levels. Quantum computing, as it matures, may benefit cognitive computing in all these aspects."