Entropica Labs: Shaping the Future of the Quantum Software Stack

Everything in nature is subject to noise, tiny fluctuations that randomly alter the state of a system and lead to errors when transmitting or processing data. 

The information stored and processed by classical computers is remarkably robust, resulting in error rates that have become negligible. In fact, the error rate in your laptop is so astonishingly low—approximately one in a quintillion—that it can be considered faultless for everyday use. 

However, when it comes to quantum computing, a different story unfolds. Quantum computers utilize qubits, the delicate building blocks of quantum information, which are highly susceptible to even the tiniest environmental disturbance. These disturbances introduce errors into quantum computations, errors that can quickly accumulate and disrupt the flow of information, posing a formidable challenge in making quantum computers useful. 

Entropica Labs was founded by Tommaso Demarie and Ewan Munro in 2018 with the ambition of creating software tools to make quantum computers useful. In May 2020, they raised 2.6m Singapore dollars (approximately US$1.8m) in seed funding, led by Elev8.vc and with the participation of SGInnovate, Wavemaker Partners, Lim Teck Lee Group, Japanese IT corporate Group TIS Inc, V1 Capital, and Entrepreneur First. Today, Entropica Labs is building the software layer to enable error correction and accelerate the arrival of useful, fault-tolerant quantum computers. 

Learn more about the future of quantum error correction from our interview with the co-founder and CEO, Tommaso Demarie: 

Why Did You Start Entropica Labs?

Quantum computing is such an exciting technology and industry. There is really nothing else I feel I should be doing—with an exception, neurosciences. Every day I am grateful for having this opportunity. I grew up with a passion for both science and science fiction, immersing myself in the works of authors like Asimov, Herbert, and Gibson—and always fascinated by artificial intelligence, robotics, and computing. 

So I decided to study physics. I did my bachelor’s and eventually a master’s in Italy at the University of Turin, during which I crossed paths with a professor in Oceanography, of all subjects, who was exploring how to apply quantum computing to his research. That’s when I also started looking into it, and it slowly became an obsession. Using the fundamental properties of nature for computing is just mind-blowing.

After one year of working at a bank in Italy—developing statistical models for risk management—I decided to pursue a Ph.D. in quantum computing at Macquarie University in Sydney, Australia. Following that, in 2014, I joined the Singapore University of Technology and Design and the Centre for Quantum Technologies (CQT) in Singapore as a postdoctoral researcher. My area of research was the theory of obfuscation and verification of quantum computations, alongside the development of protocols for measurement-based quantum computing—a framework I continue to love very much. Still, my desire was to create something that expanded beyond the boundaries of publications and academia. This aspiration gained further momentum following the early work by IBM in 2016, which demonstrated the ongoing transition of quantum computing from research to industry. 

I was determined to be part of the evolution of quantum computing by not just envisioning its potential but by rolling up my sleeves and building a company, putting my money where my mouth is. So together with Ewan, a trusted friend and colleague from my time at CQT, we recognized a critical gap in the emerging technology landscape. As numerous startups and larger companies focused on the development of quantum hardware, it became evident that without a new software stack designed specifically for quantum computing, its potential would remain untapped. Driven by this realization and by the prospect of the founders’ journey, we took the decision to leave our academic positions and co-founded Entropica Labs. 

After a false start, where we sought to leverage quantum computers for computational biology—an attempt scuttled by the realization that we were not the right team to achieve that goal—we found our path and set our sights on the big target, the big and scary target, of building software tools to enable quantum error correction and, with it, fault-tolerant quantum computing.

How Does Quantum Error Correction Work?

For quantum computers to be useful, they need to solve computations much faster than classical computers. This not only involves leveraging quantum effects to have an advantage but also requires lots and lots of qubits for error correction. The main challenge is that quantum computers are inherently imperfect. The quantum state carried by the qubits is fragile, and as the qubits interact with their environment, for example, to input and readout information or to perform computational steps, errors are introduced into the quantum computation. 

Now, how do you fix those errors? By building a software layer to execute quantum error correction codes that ensure that the quantum computation does not go off-rails. 

To give you an analogy, imagine a computation is like taking a journey. If you’re going from Italy to Germany, there are different ways you could travel: by bike, by car, or by plane. In our analogy, this is like computing on a phone, a PC, or a supercomputer. But, some destinations, such as Mars, are impossible to reach even with an airplane. So you’d need a spaceship, or in this case, a quantum computer. 

Because of errors, however, the quantum computation doesn’t go the way it’s intended to go. It’s like the spaceship drifting as far as Saturn instead of reaching Mars. So you need a control system to keep your spaceship on track—and that’s what we do, building a control system for quantum computers. 

In practice, quantum error correction consists of an active feedback process where errors are continuously detected and removed from the computation. Quantum information is made resistant to the effects of noise by spreading it across multiple qubits. Therefore, quantum computations become resilient by introducing redundancies in the process. 

There are some quantum error correction schemes, like the surface code, that have been mathematically proven to work. However, they require many more qubits than current-day quantum computers can host and only work if the error rate is low enough, which is just beyond the reach of current quantum computers.

Developing software tools to design, test and implement error correction schemes relies on our ability to use and understand the architecture, limitations and capabilities of existing quantum computing systems. To do so, we have built software tools to run hybrid quantum-classical algorithms like the quantum approximate optimization algorithm (QAOA), which can be implemented already on today’s quantum computers. This line of work is allowing us to map out the gaps in the quantum software stack, benchmark the devices, track the scale of improvements and better position ourselves to demonstrate error correction on future generations of quantum devices.

How Did You Evaluate Your Startup Idea?

The quantum computing market is still nascent, and there is uncertainty about how the industry and technology will evolve. But some problems are here to stay: suppressing errors in quantum computations is one such problem. We are building the tools to enable quantum error correction at scale, and while there is still work to do to validate our business model and reach product-market fit, our experience in the market makes us confident we are on the right track. 

The industry is entering a new phase, where we expect to see additional efforts by hardware companies to demonstrate their devices’ capabilities in terms of error correction towards fault tolerance. With larger quantum computers available, better control, and improved functionalities, our strategy is to collaborate with forward-looking hardware partners and use our software tools to demonstrate ever-increasing quantum error correction schemes leveraging cloud quantum systems. The beauty of this problem is that it applies horizontally to all hardware, offering multiple entry points for validation. We have put together a strong, dedicated team that has built extraordinary expertise in using quantum computers. Today, no one really knows how to do quantum error correction at scale. Ultimately, the industry will need a programmatic way – and that’s what we’re developing.

We’ve now been long enough in the market to get a good intuition of customer pain points moving forward. Your objective should be to genuinely understand the problem you’re solving because, in quantum computing, most players are very protective of their efforts. My business compass is: imagine tomorrow another startup shows up and offers us a software solution promising to improve our workflows, would we buy it? Our goal is to make it a no-brainer for our partners to say, yes, this is what we need. 

What Advice Would You Give Fellow Deep Tech Founders?

Focus on and have conviction in your strengths. Solve a problem that you love solving. And when you deal with venture capitalists, be aware that their job is not to be experts in the details of your technology. Occasionally, they will try to apply cookie-cutter solutions to your efforts, which can be both confusing and frustrating.

Your role as a founder is to set expectations right, guide your investors through the steps you envision to succeed, be honest about the timescale, and maintain your conviction about your way of solving the problem. I have seen countless stories of startups struggling because their founders mistake pleasing investors for progress. You should first please your customers, project partners, and employees. And if you do that right—and you are lucky—you will end up pleasing the investors too. 

Quantum computing will happen faster than most people expect, and there are a lot of open and exciting problems in the space. If you’re reading this article and you get excited about quantum computing, I can only encourage you to get your hands dirty and start working on one of these beautiful problems. Most companies and people do very established things, but in quantum computing, you have the chance to shape the future of this nascent industry. It’s like being part of the computing revolution in the 1960s. 

Further Reading

Please have a look at our open-source project OpenQAOA on GitHub, a Python library for addressing quantum optimization problems as described in this research paper. 

Singapore quantum computing startup Entropica Labs bags $1.8m in seed funding – Press release about Entropica’s seed round on techinasia.com 

Quantum computing on a budget: a practical example of cost-related trade-offs – Read more on Entropica’s blog about how to make quantum computing economic

BMW Chooses Honeywell Model H1 Quantum Computer And Entropica Labs For Supply Chain Quantum Proof-Of-Concept – Forbes article about Entropica’s pilot with BMW and Honeywell.

Implementing real-world optimisation use cases in state-of-the-art quantum devices – Read more about solving optimization problems on Entropica’s Medium blog

Unbalanced penalization: A new approach to encode inequality constraints of combinatorial problems for quantum optimization algorithms – Arxiv paper about formulating and solving hard optimization problems by Entropica. 

The Houdayer Algorithm: Overview, Extensions, and Applications – Learn more about Entropica’s work on cluster Monte-Carlo methods, a family of powerful algorithms to solve optimization problems, in this arXiv paper