Rigetti Partners with CDL to Drive Quantum Machine Learning

by Madhav Thattai and Will Zeng, Rigetti Computing

We are excited to announce a partnership between Rigetti Computing and the Creative Destruction Lab (CDL) to drive the development of quantum machine learning and help realize its full potential.

Quantum computers will transform many fields, and the impact on optimization and machine learning (ML) will be among the most profound. With the rise of GPU computing and ML-specific processor architectures, the world has seen hardware-accelerated machine learning revolutionize many technologies and products.

Yet this is just a glimpse of what’s to come. There are large classes of machine learning problems that are inaccessible to the best of today’s hardware. The exponential complexity of these problems requires a resource that can scale to match their needs. Quantum computing is the only technology that has this capability. It will drive significant advances in machine learning [1], and a new ecosystem of quantum software engineers and application developers is now starting to create the first building blocks.

So far, development has lacked general-purpose quantum hardware and a real programming stack for developers and aspiring companies to use — critical requirements to accelerate building applications. Our partnership with CDL and their companies is aimed at changing that.

CDL’s first cohort of quantum machine learning startups will use our programming environment and toolkit, Forest, to develop applications for quantum hardware over the cloud. Forest gives these startups a big leg up because we built it around quantum/classical hybrid programming, a new paradigm which will become the dominant form of quantum programming.

In the past [2], quantum algorithms were designed for large and perfect quantum computers and typically ran exclusively on a quantum processor. Quantum/classical hybrid programming changes this, using classical computers in conjunction with quantum processors. In this approach, the quantum resource is focused on only the parts of an algorithm where it can provide a huge advantage — and the traditional computer does the rest. This drastically reduces the performance requirement on the quantum side, allowing developers to do much more with early hardware generations.

We knew this programming paradigm would be important, and that is why we specifically built Forest with a focus on the hybrid architecture required for near-term applications [3]. Forest provides the programming environment that developers — like those in the CDL program — need, and also gives them access to the superconducting quantum processors that we have built at Rigetti. Working directly with quantum hardware is critical for anyone looking to exploit the power of these first machines.

CDL shares our vision for how to invest in and build the quantum software industry, and has shown real leadership in implementing that vision. We are incredibly excited to support their quantum machine learning startups in discovering and developing applications using Forest. These companies will impact the world in ways we cannot imagine today.

[1] Biamonte et al., Quantum Machine Learning https://arxiv.org/abs/1611.09347

[2] P. Shor, Polynomial-Time Algorithms for Prime Factorization and Discrete Logarithms on a Quantum Computer https://arxiv.org/abs/quant-ph/9508027

[3] R. Smith, M. J. Curtis, W. J. Zeng, A Practical Quantum Instruction Set Architecture https://arxiv.org/abs/1608.03355