Zapata Computing has recently been making headlines with new access to its Orquestra service and availability across a number of platforms. Recently Zapata also re-shuffled its scientific advisory board to include notable figures in the Quantum Community. We ask Zapata about the changes happening and also about the future of Quantum Computing.
In this interview we talk about the marketplace, technologies, Quantum Machine Learning, Quantum Operating Systems, with Yudong Cao, CTO and Founder and Zapata Computing and Peter Love, Associate Professor of Physics and Astronomy at Tufts University.
QZ: There are quite a few luminary members of the new scientific advisory board. How are you positioning the focus of the board?
The function of luminaries is to shed light, and we hope to focus the illumination provided by the scientific advisory board on the deep scientific questions at the core of Zapata’s mission. Can we realize quantum advantage on near term devices? If so, for which problems? What are the cutting edge research developments that Zapata should pay attention to? Where can we best focus Zapata’s research team to make the biggest contribution to advancing the field?
The Zapata Scientific Advisory Board (SAB)
The founding members include Andrew Childs, Professor in the Department of Computer Science and Institute for Advanced Computer Studies at the University of Maryland; Jens Eisert, Professor of Quantum Physics at Freie Universität Berlin; Aram Harrow, Associate Professor of Physics at the MIT Center for Theoretical Physics; Arthur Jaffe, Landon T. Clay Professor of Mathematics and Theoretical Science at Harvard; Peter Love, Associate Professor of Physics and Astronomy at Tufts University; Will Oliver, Associate Professor of Electrical Engineering and Computer Science and Lincoln Laboratory Fellow at MIT, Director of the MIT Center for Quantum Engineering, and Associate Director of the MIT Research Laboratory of Electronics; and James Whitfield, Assistant Professor of Physics, Dartmouth.
QZ: Orquestra is your quantum platform. Many in the community will be perhaps more familiar with IBM Q and there is growing interest in the platforms and perhaps a fight for this ground, rather like operating systems. Why should anyone use Orquestra rather than using hardware platforms directly?
First of all, Orquestra is not a quantum programming library that displaces Qiskit or any other software libraries. Instead, it is a software tool occupying a different layer of abstraction from these libraries. One can very well use Orquestra with Qiskit, cirq, pyquil or any combination of quantum programming library as the underlying components. An important part of Orquestra is a workflow management system that allows interoperability between different paradigms of quantum as well as classical programming. This interoperability allows Orquestra to help people use multiple quantum resources, including multiple hardware platforms and software libraries, and also classical HPC resources in concert with each other.
QZ: Tell us about your research? What are you working on?
Zapata is focused right now on exploring quantum advantage for near-term noisy devices. Zapata is very interested in both bringing existing quantum techniques such as VQE, QAOA and QML into the set of resources available for Orquestra users and also developing new tools and techniques to improve the existing approaches. There are video discussions of our recent papers on the Youtube channel.
QZ: Where do you see the future of the quantum operating system? Will we see just a few operating systems to manage heterogeneous hardware?
“Quantum operating system” is a term that is rather ambiguous considering how “operating system” carries a specific set of meanings in the classical world. Our preferred term for describing Orquestra is “quantum operating environment”, which translates to a set of software features and related infrastructure which enables seamless development and deployment of hybrid quantum-classical information processing. Such quantum-classical interoperability is also extensible to other exotic substrates for computation that may make up a heterogeneous computing stack.
QZ: At the moment there is a lot of interest in QC with many companies exploring, but it would appear there is no killer application, what do you think will become the killer application?, i.e. industry or process? VQE in drug discovery? Optimization? A new algorithm?
The question of “killer application” is tied to the question of “what would be the first instance of quantum advantage”, where the definition of “quantum advantage” is still a heatedly debated subject. There are quantum algorithms such that if we have a large-scale error corrected machine in ten years or so those algorithms will have a large impact. In the near term, for noisy devices I agree things are much less clear. Optimization is terribly difficult in the worst case and so far there is not a clear quantum advantage for any problem, but because of the ubiquity of optimization problems we should keep looking!
QZ: Many of our readers are keen to explore Quantum Computing, but are perhaps wondering the best way to get started? Is there anything to convince them that Quantum is likely to disrupt their industry?
This requires serious study for each possible application area and the answer will differ a lot by sector. Sometimes the answer will just be – yes, quantum computing is very exciting, but not for your industry. Zapata has engaged in a number of such studies for various clients so the simple answer is that they should call Zapata!
QZ: Where are we in the evolution of Quantum Computing? We like to think of historical terms such as microprocessor history. What are we in that evolution? We have a variety of hardware (qubits) and software, but would love to hear where you think we are on the technology pathway.
I think we are trying to build the quantum equivalent of the first large scale vacuum tube computers, such as ENIAC. Very large, not very powerful compared to what is coming later, but a huge leap from the mechanical calculators that preceded them.
QZ: Quantum Machine Learning is a term that has been bandied around. Right now, it is early days, what do you see as the future of QML?
I think QML struggles now because most data is classical. To use QML you must first load the classical data to quantum memory – that’s hard. As we get more quantum data from quantum sensors and quantum simulation I think QML will come into its own. Of course, QML techniques are fascinating as quantum algorithms in their own right. There are specific approaches that we are developing at Zapata which will enable us to address real-world machine learning tasks with near-term quantum devices.
QZ: What is next for Orquestra? Can you give us a sneak peak into where you’re taking the development of the platform?
Orquestra is evolving to include more features across ML, chemistry and optimization, as well as a better user interface and data management features. We are starting to see our users creating tasks for others to use. As always, we’ll be committed to adding hardware and software integrations and more compatibility so Orquestra can continue to be extremely flexible, modular and powerful for academic and enterprise users.
Thank you to the Zapata team to talking to Quantum Zeitgeist!
