As the global head of D-Wave’s Technical Advising team, I spend much of my time thinking about how to best support our users and customers, especially when it comes to trying our products for the first time or building their first application or code example. Over my past six years in this space, I’ve observed two key issues that often come up.
First, technical users are eager to dive right into their quantum exploration without taking the time to set up a plan for the project.
Second, technical users are not sure what they need to show their decision makers to move forward with a quantum solution.
Consider your first time trying out a new product or tool. How did you test it? What was the process? How could you tell if it worked for you? And how did you decide if you should continue to use it? These types of questions are all relevant with quantum computing too.
I see users building applications and testing out a new quantum demo without considering what they are looking for at the end. Imagine trying out a new pair of running shoes without knowing anything about the racecourse or the terrain. Are you running for speed or distance? Is the course flat or through the mountains? A successful trial of your new shoes would have a very different plan and very different success metrics for each of these scenarios. Approaching your shoe trial without understanding the end goals or requirements will make for an interesting race day.
For any quantum computing effort, project planning is crucial. Take some time to plan out your project, examine your success criteria, and consider how to interpret your results right at the start of your project. These can dramatically improve the chances of a successful trial.
Here’s our framework to help users drive a successful quantum initiative:
D-Wave’s Project Planning Framework
1. Examine your use case.
Choosing a good use case and clearly defining that use case can be more challenging than it seems.
First, choose a use case that is relevant. Are you hoping to prove the value of quantum computing within your business unit or to your manager? Choose a use case that they care about and that has real business value.
Second, choose a use case that is easy to start with. Are there existing code repositories or papers that you can leverage to get started? Often you don’t need to start from scratch. Instead, you can build off work from others.
Tip: Take a look at our use case library or code examples for inspiration.
Third, choose a use case that is a good fit for D-Wave’s technology. As a new user, this part can be tricky. D-Wave provides a lot of useful resources to guide you through this process, including webinars that explain how to identify relevant and promising use cases — I’ve linked one below. Consider things like the scale of the problem in the real world: will our solvers be able to tackle a full-scale problem, or will you have to consider simplified or reduced-size instances? If the problem needs to be reduced, how will you explain the value of the solution to your decision maker?
Watch our webinar on Building a Quantum Hybrid Application
2. Determine the solution requirements and success criteria.
We are sometimes so eager to start formulating and coding, that we forget to consider what we are trying to achieve. It’s important to consider not only what would be impressive to your business manager, but also what’s realistic in the given time frame you have for your trial. In general, success criteria should be able to be confirmed with a simple yes or no question. For example, “did my solution run faster than XX?” or “is my solution able to solve a problem of size YY?”. Determining these criteria up front with your team can make the job of showing your work at the end much easier — you will know clearly whether you achieved what is needed to move forward with quantum annealing.
Some suggestions to consider:
Basic implementation. If you’re simply trying to answer the question “can I build a simple example using these products?” then you can measure your success by a “yes” or “no.” However, you might also consider adding things like visualizations or demos to your example to demonstrate the potential to your less-technical peers.
Run time. We hear customers say that they want their D-Wave solution to run faster. The important follow-up questions that we always ask are “faster than what?” and “how much faster?”. If you’re looking to demonstrate a solution with a faster run time, then you’ll need to have existing classical solutions handy for comparison. Keep in mind that Python is not the most efficient programming language — having two programmers develop code to complete the same task can result in drastically different run times!
Last, but definitely not least, we usually only start to see run time advantage on problems that are large enough. D-Wave’s hybrid solvers take a few seconds to run — if your problem can run classically faster than that, then don’t expect to see a faster quantum computing or hybrid quantum solution. In general, comparing a cloud-based quantum solution against a local classical solution on a small problem likely will not show that quantum is faster. Scaling to larger problems where classical solvers start to slow down is where we can see quantum benefits.
Solution quality. We see solution quality benefits in quantum applications. If you don’t care about run time, or you don’t have an existing classical solution, then this might be a good success metric to consider. Still, you’ll want to have solution benchmarks available for some hard problem instances where you believe improvements are possible to compare against. These solutions may be derived manually or with a classical method. Once again, important questions to ask are “better than what?” and “how much better?”
Solution scale. A good metric to consider for a challenging problem is solution scale. Can your problem formulation and Python program handle a problem at a scale that is needed by your business? If this is your metric, keep in mind that different problem formulations can scale differently, so you will want to consider several variations.
3. Outline your plan for success.
In advising our users through this initial exploration of D-Wave’s technologies, we’ve found it useful to help outline a plan that guides them through their trial. This way we can ensure that we have a plan in place to achieve the goals within the given timeframe. These trials can be quite short — it might be a Quickstart training package (1 month of access) or someone with a deadline from their manager.
To facilitate this planning process, we start with the following outline where we fill in the timeframe column with internal deadlines to stay on track. Further detail on many of these steps can be found in our Workflow for Developing Applications.
As you may be able to see from the description and deliverables column, determining your success criteria and metrics first is particularly important and can be done even before you open an account on our platform. These decisions will dictate how you approach each of the steps in this outline. By working through these steps, you can develop not only an initial application using D-Wave technology, but also determine if your success criteria are satisfied. This, in turn, can mitigate any unexpected questions about the success of your trial as you present your results to your decision makers.
Want to learn more? Reach out to us at sales@dwavesys.com and let us work with you on your project plan.
About D-Wave
D-Wave is a leader in the development and delivery of quantum computing systems, software, and services, and is the world’s first commercial supplier of quantum computers. Our mission is to unlock the power of quantum computing today to benefit business and society. We do this by delivering customer value with practical quantum applications for problems as diverse as logistics, artificial intelligence, materials sciences, drug discovery, scheduling, cybersecurity, fault detection, and financial modeling.
Discover more at dwavequantum.com and begin your quantum journey today.