Gunnar Assmy - Fotolia
CES 2020: IBM grows Q network collaborators
ExxonMobil hopes quantum computing will enable it to model processes at the atomic scale – something that cannot be achieved today
During a presentation at CES in Las Vegas, Dario Gil, director of IBM Research, urged developers to start thinking about how to use quantum computing to solve "exponentially" complex problems.
“We are in the phase of getting the world quantum ready, he said during a presentation exploring the quantum decade during CES 2020. “A whole generation of developers will need to learn how to program quantum computers.”
In the last 50 to 60 years, Gil said the industry has built an information technology edifice in what is now known as classical information theory, which decouples zeros and ones from the physical implementation. He said: "It is this separation that allowed us to come to understand that both punch cards and DNA can be understood as carriers of information.” Gil said this theory combined with Moore’s Law has made bits ubiquitous and almost free.
But he said there is now another way to understand information. “The building block of the world of information is not the bit, the zero and one, but something known as the quantum bit or qubit.” There are three ideas that form the backbone of quantum computing: superposition, interference and entanglement. He described superposition as a bit like the state of a coin when it is spinning. It is neither “heads” nor "tails" but a combination of the two. “In classical computing, if I spun two coins the probability of getting a head or a tail on either coing is independent.
The idea of entangled qubits in the quantum world means that there can be situations where if one coin is heads, the other will always be heads, and if it is tails the other will always be tails. This means the two coins are entangled such that it is impossible to measure their state (either heads or tails) independently. Gil said: “This property is weird in its nature but reflects the nature of physical reality. You can have known local action in the Universe.”
He said there is a relationship between quantum entanglement and the world of information. “If I had a hundred perfect qubits that were entangled and tried to represent all the possible states of that entanglement using zeros and ones, I would have to devote every atom on Earth to store [the data]. A quantum computer with 280 qubits would need every atom in the known universe.”
The third property of quantum computing is interference, Gil added. “We see this interference ocean waves. You can have waves that interfere with one another.” These can combine to form peaks and troughs.
Unlike a classical computer, which is programmed in zeros and ones, Gil said the first task in programming a quantum computer is to put the machine in a superposition state. A two bit qubit has 22 states, ie four possible states. If these states are represented as dots on a sphere, a one could be represented as a dot on the North Pole, zero, a dot on the South Pole, and a dot at the equator would be used to represent a state that is both zero and one. Each dot has two values to pinpoint its location in the sphere. To put information into a quantum computer, involves changing where the dot is located within the sphere.
Theoretically, programming a quantum computer involves taking the spheres holding information and interfere them with one another in order to maximise correct answers, and minimise incorrect answers. “Like waves in an ocean, I can explore a series of possibilities to find the answers I want, with this process of interference,” said Gil.
Why quantum matters
Gil said classical computers are good at solving easy problems, where the number of variables to solve are not exponential. He said: "There are other problems that are exponentially more difficult to solve, because the number of variables are exponential in nature.” Examples include modeling nature. The process of [modeling] chemicals and materials and understanding how nature behaves is exponential.” Running accurate models of the physical world is too complex for traditional computing.
For the last few years IBM Q, the company’s quantum computer, has been accessible via the cloud. According to Gil, more than 200,000 users are now on using IBM Q experience initiative.
IBM wants to encourage organisations to experiment with quantum computing and develop application areas, so it has developed software development kits to enable programmers to experiment with programming for quantum computing. The IBM Q System One can be programmed using the Quiskit open source software development library.
Read more about quantum computing
- IBM’s Bob Sutor discusses Big Blue’s new quantum systems and computation centre, the realities of quantum computing today and how IT pros will code for these systems in the future.
- Google claims it has developed an algorithm for a quantum computer that would take a traditional “classical” computer 10,000 years to run. We investigate.
ExxonMobil is the one of the business collaborating in the IBM Q network at the start of 2019.
Vijay Swarup, vice-president of research and development at ExxonMobil, said that quantum computing is needed because the complexity of the energy sector’s computing needs is increasing.
ExxonMobil wants to be able to use quantum computing to help it build more accurate models and simulation, said Swarup. “We want to be able to understand energy at its most fundamental level, at the atomic scale, which we cannot model today. We have predictors, but we don’t actually know their accuracy.”
One possible application area for quantum computing in the energy sector is carbon capture, which requires the development of new materials that can capture carbon as efficiently as possible. Improving the accuracy of the theoretical model of the material using quantum computing could enable better carbon capture material to be developed.
To support this sustainably, he said it is necessary to understand the chemical reactions that take place in reactors that convert one substance into another. “We need to understand what is happening at the atomic scale to make models – not just predictors, but accurate models. This would be a huge breakthrough.”