D-Wave Systems have been at the forefront of the quantum computing space since their first commercial quantum computer was sold in 2011. Now with the largest and most advanced machine to date, the 4th Generation 2000Q, D-Wave are empowering global organisations with the tools to tackle the most challenging computational problems.
This week, Quantum Business spoke with Alan Baratz, the senior VP of software and applications at D-Wave Systems. Baratz runs D-Wave’s software side and is responsible for the development of tools, algorithm research and application of software development. With 30 years at the top of the technology space, Baratz has seen new forms of computation evolve. He told us that today he feels fortunate to be at the helm of a radical organisation making a real impact on the world.
“In terms of my interest in quantum. I’ve been interested in new approaches to computing since my time at MIT. When I was an undergraduate, neural networks were an interesting curiosity. There was some early research around that. Those were the early days thinking about different computing architectures. My work was focused on the design and analysis of algorithms and complexity theory. I was always intrigued by the notion of potentially being able to develop different types of computing architectures that might be able to attack very difficult problems – NP Hard Problems.”
Tackling NP Hard problems
NP Hardness are a class of problems for which nobody to date has been able to develop efficient algorithms on current computing architectures. According to Baratz, these are problems which have been proven to be equivalent in the sense that if you can solve any one of them efficiently then you can solve all of them efficiently. A frequently used example of NP-Hardness is the ‘Travelling Salesman Problem’ which asks ‘Given a list of cities and the distances between each pair of cities, what is the shortest possible route that visits each city and returns to the origin city?’
“NP Hardness are thought to be the most computationally difficult problems in the world. It’s now been many years since this theory was first defined. Many problems are now known to be NP hard – amongst the most computationally difficult and we still don’t have efficient algorithms for solving them. But there’s always a search for being able to come up with software that will solve them. Now the search is for new types of computing architectures.”
Where Does Quantum Computing Come In?
“Quantum Computing is one of those approaches to computing that is very different from how we’re doing computing today. It has the potential to solve very difficult problems efficiently. Closely related to this but not known to be NP hard, is the problem of factoring large semi-primes, large numbers that are a product of two prime numbers.”
Baratz explains that like NP hard problems, this is a challenge that has been around for a long period of time and nobody has been able to come up with efficient algorithms for doing it. “All of our cryptography algorithms today are based on that. In the same way we look for new kinds of computing architectures to solve NP hard problems, we’re always interested for new computing architectures that can factor numbers. As a result, this forces us to be thinking of even stronger cryptography algorithms.”
“Quantum Computing is one of those new kinds of computing architectures that holds the potential to attack these very difficult problems. For me that’s very interesting and I’m thrilled to have an opportunity to be a part of figuring out how to harness that power.”
Quantum Computing Market
Today there is more and more capital flooding into global research hubs, start-ups, and large organisations. The focus is on bringing commercial quantum computing to the masses. According to the Economist, total public funding will amount to $1500 million over the next several years. Analysts predict that quantum computing could hold the key to doubling the high-end computing market from $5 billion to $10 billion over the next decade. But what is the commercial situation today?
“When you started you said you were interested in the quantum computing market. I was a bit amused when you used the term market – in the sense that there’s really only one company in the market today that has a quantum computer that is not only available for sale but has been sold. That is D-Wave.”
D-Wave sold the first 2000Q quantum computer in January 2017 to the cybersecurity firm, Temporal Defence Stystems. The 2000 quantum bit (qubit) machine was a follow up to the 1000 qubit 2x, released in August 2015. Baratz explained more about the amazing progress being done at D-Wave and what sets this revolutionary machine apart from the rest.
“The 2000Q is our most recent model and it’s our 4th generation system. Nobody else is pursuing either the same architecture as us and nobody has a commercially available quantum computer. The market is really a company of one right now. There’s a lot of research going on and a lot of rhetoric but the market is only a market of one right now.”
How Does It Work?
There are many different approaches to building a quantum computer and D-Wave’s approach is known as quantum annealing. Baratz explains the similarities and differences across the space.
“When we’re talking about individual qubits, there’s really no difference between what we do and the approaches taken elsewhere. A qubit is simply a unit of memory like a bit in a traditional computer, except for the fact that a qubit can be both 0 and 1 at the same time. Because qubits can be both units simultaneously, a quantum computer has the ability to explore many many different possible solutions to a problem at a given point in time. When it comes to the core element of the qubit. There’s not a lot of difference between the underlying technology for creating a qubit – whether you’re using an annealing architecture or a gate model architecture. The difference comes in how those qubits actually get used.”
“In the annealing model we are using the collection of qubits to explore a very complicated landscape and look for the lowest point within that landscape. At it’s core that is what the D-Wave quantum computer does. That is what annealing is all about – It is all about finding the lowest point in a very complicated, multi-dimensional landscape.”
“If you wanted to think about this in the context of qubits, you have all these qubits that can be in a 0 and 1 state simultaneously. If you were to have a specific value for any one qubit, a collection of 0s and 1s across all the qubits that would be one point in the energy landscape. Because these qubit can be 0 and 1 at the same time in some sense the computer is seeing all the points in the energy landscape all at the same time. And settle into the set of qubits that represent the lowest point in the energy landscape. Again, this is really about how we’re using the qubits more than how the individual qubit is built.”
What is the Power?
According to Baratz, a very broad collection of problems can be formulated as the problem of finding the lowest point in a complicated landscape. “Many of the NP problems, can be reformulated as an energy landscape where the optimal solution is the lowest point. The annealing approach can be used in a broad array of optimisation problems, machine learning problems, material simulation problems – they can all be backed into this problem of finding a low-point in a complicated landscape. This is what annealing is about – finding a low point in a complex landscape.”
Who Are the Users of D-Wave?
D-Wave’s customers today are global leaders across an array of industry verticals. This is a network organisations that has grown. “Google NASA is one of our users, Lockheed is another customer, Los Alamos Labs, Oak Ridge National Labs is another customer. We have a collection of other customers around the world. And the truth of the matter is that these customers are using our system to solve a very diverse array of different types of problems.”
“It is everything from material simulation problems, which is an example of the type of thing Google and Nasa are working on to Machine Learning problems, which Los Alamos and a number of our customers around the world are working on to constraint satisfaction problems, like Lockheed and Los Alamos and Ockridge to Optimisation problems.”
“A good example (optimisation) is the Volkswagen Application. Basically Volkswagen use the D-Wave system to optimise taxi cab routing in a very large congested city. The city was Beijing. And what they were doing was looking at the optimal way to route taxis so that everybody gets a reasonably kind and affordable ride. This is a large, complex optimisation problem that VW used our 2000Q to solve.”
There’s a broad array of other problems as well in the optimisation space. “One of our customers wanted to extract better data from polling. The aim was to attain more complex relationships in the polling data – so how much better could you do at predicting results? Essentially what they did was they went back and took polling data from the last presidential election where all the polls were predicting a landslide Hilary victory. What they did was they started looking at additional correlations including correlations across states rather than individual polls within states. Through that analysis they came out concluding that trump would win. Essentially being able to correlate more data is a much more complicated optimisation problem and they use that system for that analysis.”
“The system has been used for protein folding and drug discovery area. It has been used to do fault detection in circuits. So if you’re doing circuit design work: Maybe you’ve got a complex circuit that you’ve designed, you put some inputs on the circuits and you read the outputs and the outputs are wrong but you don’t know where the fault is in the circuit. Our system can tell you which part of the system has faults.”
Baratz outlined the major areas where quantum computers can advance humanity over the next decade. “My mind first goes to medicine, drug discovery and pharma. We do have customers doing work in that arena. At the core of that is protein folding, being able to understand how proteins actually live within the body and use that to create better drugs. It also goes well beyond that. For example, we’re doing some work now for DNA markers for certain types of cancer, Looking at large amounts of data and looking at correlations in it.”
“We talked about traffic routing in Beijing. Now think of that as a component of self driving cars. The ability for cars to be able to route themselves in real time in a very efficient way. As a result get the traffic nightmare out of our daily commute. There are just a variety of ways that quantum computers can add value to our daily lives.”
What are you looking forward to in the next 6 months?
“There’s one really important and exciting thing I’m personally involved with right now. Up until now the primary way that customers got access to our system was by purchasing a system or by purchasing a large block of time on a system. One of the things that we are working through right now is how to make our our systems more broadly to not only a wider class of customers around the world but also to developers who may be interested in learning about quantum and how to develop applications for quantum, helping to identify compelling applications for the system. One of the things that we’re working on now is how to take a collection of our systems and open them up more broadly to more customers and people.”
Dr Alan Baratz is the SVP of Software and Applications at D-Wave Systems
Welcome to Quantumbusiness.org! The first news portal dedicated to exploring the quantum computing revolution and its forthcoming impact on global industry. For more information on content creation and the opportunity to share your story with the world, please contact our lead editor Hal Briggs [firstname.lastname@example.org].
Written by Hal Briggs from Quantum Business