Implementing effective error mitigation and correction is a critical next step in advancing quantum computing. While a lot of attention has been given to efforts to improve the underlying ‘noisy’ hardware, there’s been, perhaps, less spotlighting of similar efforts to develop software strategies for coping with these errors. That’s now changing and the emerging software approaches show great promise. Q-Ctrl is one of a growing number of companies tackling this problem. It boasts several successful engagements including with one with IBM to improve gate fidelity and error reduction rates.
“The question that we asked ourselves is how do you accelerate the whole industry? One way to do that is by starting with this thesis saying, how can software make quantum technology more useful. So, imagine you take any quantum architecture – superconducting qubits, trapped ion, whatever – and you put a piece of software on it and the hardware can now suddenly become 100x or 1000x better. That is the problem that we sought to solve,” said Aravind Ratnam, who was named Q-Ctrl’s chief strategy officer yesterday.
Think of Q-Ctrl as a middleware/firmware company, at least in terms of the quantum computing part of its business. Q-Ctrl also has distinct efforts in quantum sensing (hardware and software), broader quantum engineering (control systems), and professional services. But its core quantum computing products are around this idea of creating software to mitigate error and optimize performance on diverse quantum hardware platforms.
“We have the notion of an abstraction layer that allows us to work almost equally well on all of the [quantum computing] architectures simultaneously and we have published papers to this effect. The dream is this idea to have a common software layer, have software libraries that are optimized to certain architectures, and allow users to download it and make their technology instantly useful or instantly more useful than it was before,” Ratnam told Technovanguard.
Founded in 2017 and based in Australia, Q-Ctrl is led by CEO Michael Biercuk, a professor of physics and quantum technology, University of Sydney, and a chief investigator at the ARC Centre of Excellence for Engineering Quantum Systems. Ratnam is the “first external hire into the C-Suite since the company was founded,” reported Q-Ctrl, which touts Ratnam as having “a wealth of experience from Silicon Valley startups where he led product, technology and corporate strategy. Most recently, Ratnam served as head of products for Sense Photonics (acquired by NASDAQ: OUST) and vice president of products for AEye Inc.”
He’s charged with developing and leading a strategy for 10x growth over the next three years. (No pressure.) Here’s his brief description of two main business thrusts:
- Quantum sensing. “I would almost regard that separately (from quantum computing), because there’s all these different applications that can benefit from added sensitivity to measurements, whether it’s navigation, topography, magnetometry, gravimetry, etc. This [involves] a combination of hardware and software. We have several contracts, some of which we have announced, including with the Australian Space Agency, and others that we have not introduced. The quantum sensing effort is going on in parallel; we are building out hardware that is specifically for quantum sensing. Eventually, what’s going to happen is that this hardware will either work in conjunction with what customers have or get embedded in [systems],” said Ratnam.
- Quantum computing. “The quantum computing development and engineering effort is different. We have [currently] three products in total. One, called Border Opal, that works directly with the hardware. [It] generates very a deep level of optimization [and is] able to very precisely control what the expected error rates are and [allows you] to actually do something useful with your hardware. On top of that is another product called Aspire Opal that has automation layers,” said Ratnam. The idea here is, “once you’ve got the hardware working, it supplies the middleware layers [you] need for the algorithm to work efficiently. Think of it as almost a compiler-equivalent. We also provide a lot of other middleware that’s needed.” Q-Ctrl also offers Black Opal, a combination education/sandbox tool for learning about quantum computing.
Shown below are figures from a 2020 white paper by Q-Ctrl (Software tools for quantum control: Improving quantum computer performance through noise and error suppression) broadly describing the Q-Ctrl approach.
The company has had many engagements including work done with IBM in which Q-Ctrl reported achieving 10x improvements on logic error rates, hardware performance across the chip, and hardware stability over time. An important point here is that efforts by Q-Ctrl and other similar companies are deeply involved in understanding and characterizing underlying quantum hardware and developing error containment strategies. The reach of middleware in this early stage of quantum computing is extensive.
On its website, Q-Ctrl describes some of the IBM work thusly: “How do you decide what the pulses should look like? You start by analyzing the Hamiltonian — the energy equation — that describes the quantum computer’s operations at a physical level. This information allows you to craft hardware-aware logic gates, like the X or the Hadamard gate, specially tuned to the device being controlled, even incorporating smoothing filters that correct for distortions introduced by the dilution refrigerator’s finite bandwidth limitation.”
Q-Ctrl has been part of IBM’s Q Network since 2018. Asked if its technology was currently embedded in any IBM quantum system, Q-Ctrl supplied the statement below shortly after this article was first published.
“Q-CTRL solutions run on any hardware backend and recent tests have demonstrated up to 1000X improvement in the performance of algorithms on real hardware. IBM provides a backend and programming framework in python that allows not only the execution of algorithms, but also the definition of the individual gates programmed at the “analog” layer. Q-CTRL was happy to work with the IBM teamin defining this programming standard. Q-CTRL’s tools connect via the cloud and can be executed by any user in order to create new optimized error-robust gates, perform error-robust compilation, and even improve the measurement procedures on IBM hardware. We’re excited to have been a partner with IBM since 2018 and look forward to an expansion of joint offerings in the future.”
(For a look at IBM’s progress and its claim to three nines (99.9 percent gate fidelity) performance see Technovanguard’s recent article, IBM Breaks 100-Qubit QPU Barrier, Marks Milestones on Ambitious Roadmap.)
Asked about indicator milestones to watch for given the high expectations being set for Q-Ctrl, Ratnam said, “Product traction as I indicated earlier. We’re looking at [some] big product launches this year. There’s enterprise license growth, major [product] refinements and new customers. One of the things that is less known is that very few people in quantum are actually producing revenue, and we are one of the those [who are]. We have revenue which is growing and a substantial amount of that revenue is recurring.
“The market that we play in is a two-sided market. One market is working closely with the hardware vendors. We want to enable them. We want their hardware to work better and we want them to succeed because at the end of the day our success hinges on their success as well. The second piece is we are doing our own market development, talking to end-users in aerospace and defense, across finance and many other areas. What we’re doing [there] is two things. One is educating end-users on what quantum can do when exploring new use cases and [we’re] doing joint R&D efforts. Frequently in these use cases, when we work with an end customer, we will bring in a hardware supplier. At this time, we are working on an equal footing for everyone; we’ll see how or if that changes in the future,” he said.
It will be interesting to watch the development of middleware technologies and companies moving forward, particularly given the diversity of underlying qubit technologies.
Links to relevant papers by Q-Ctrl:
Q-CTRL demonstrations of robust quantum logic gates 10X better than default gates on IBM hardware (Published Physical Review Applied)
Q-CTRL development of deep learning to autonomously design quantum logic better than the best human-defined gates. (Published PRX Quantum)
The post Q-Ctrl – Tackling Quantum Hardware’s Noise Problems with Software appeared first on Technovanguard.