Classiq Researchers Report on a Quantum Software Development Approach to Increase Efficiency and Scalability

Insider Brief

• In a study, Classiq scientists show how EDA-inspired method reduces qubit and two-qubit gate requirements by orders of magnitude, a step the team writes could make complex quantum programs scalable for real-world applications.
• The synthesis engine dynamically optimizes quantum algorithms by abstracting high-level functional models, enabling flexibility across diverse hardware configurations and constraints.
• By integrating advanced resource management and adaptability, this approach could significantly accelerate quantum computing’s impact in fields like chemistry, finance, and machine learning.

In a recent news release, Classiq Technologies, Deloitte Tohmatsu Group, and Mitsubishi Chemical Corporation recently shared how they are leveraging quantum computing to accelerate the development of advanced materials, such as new organic electroluminescent (EL) materials. The effort was focused on streamlining quantum algorithms to enhance calculation speed.

A recent study posted on the pre-print server arXiv by researchers from Classiq may help give a bit more in-depth look at this how this method for developing quantum software actually works. By pulling in principles from electronic design automation (EDA) — a suite of software tools that enable designers to create high-level circuit descriptions automatically optimized for physical implementation — the study outlines how a similar scalable approach can significantly reduces computational resource requirements for complex quantum programs.

Similarly, in the paper, the researchers report that their method is based on separating the abstract design of quantum algorithms from their physical implementation, which essentially means developing quantum algorithms at a high level to allow software to adapt to different physical systems automatically.

According to the team, this method outperforms several current tools available in their testing. They reported orders-of-magnitude improvements in reducing computational requirements, such as the number of qubits and two-qubit gates, and showed how their approach scales for large, real-world algorithms. If it holds, this would mark a significant step forward in bridging the gap between quantum hardware advancements and the software required to harness their potential.

The study argues that existing methods, which map high-level descriptions to specific hardware, is too rigid and cannot efficiently handle the increasing complexity of quantum programs. Their approach, on the other hand, dynamically adapts to hardware constraints and user-defined optimization goals, such as minimizing gate usage or circuit width.

High-Level Functional Models

Classiq’s method leverages high-level functional models to describe the desired behavior of a quantum program, including constraints and optimization criteria. In this approach, to put it simply, programmers set goals and rules, like minimizing the use of resources, while leaving the complex work of figuring out the details to the software. It’s like giving an architect a general idea for a building and letting them handle the engineering specifics.

One example explored was a quantum walk algorithm, which, according to the paper, is a fundamental routine in quantum computing used in search and simulation tasks. The study showed how Classiq’s synthesis engine dynamically chose between different implementations of the same function, optimizing resource allocation such as auxiliary qubits. This flexibility resulted in significantly lower gate counts compared to existing platforms.

In another case, the researchers applied their approach to a Quantum Singular Value Transformation (QSVT), a method used in quantum machine learning and simulation. By helping select efficient building blocks, the synthesis engine achieved resource savings that would be impractical using manual design or existing compilers.

How Can It Be Used?

This development could be a huge help for the quantum software landscape, the team suggests. As quantum hardware grows more powerful, software capable of matching that complexity becomes essential. The EDA-inspired method presented in this study would pave the way for more practical applications of quantum computing in fields like chemistry, finance and machine learning.

The team goes farther, saying that without tools like this, the ability to surpass classical computers will be unlikely.

“Building powerful and scalable software technologies for quantum algorithm design is an essential step toward any meaningful quantum advantage,” the researchers wrote.

By improving the ability of software to adapt to diverse hardware configurations and constraints, the approach also enhances accessibility, making quantum computing more usable for developers in various industries. This could lower barriers to entry for companies looking to explore quantum technology.

Limitations

The study notes that several challenges remain.

“While we laid the foundations of the EDA-based synthesis concept and concretized it by an industry-grade software platform, there is still much to be done,” the researchers write.

Among those next steps, the team indicates that they will need to fully address how their system manages accuracy requirements in noisy intermediate-scale quantum (NISQ) devices, where balancing precision and error rates will be a critical next step. While the study focused on optimizing for specific criteria like gate count or circuit width, it did not explore co-optimizing software and error-correction protocols — a crucial aspect of future quantum computing systems.

Another limitation is the reliance on pre-existing libraries for function implementations. Although these libraries enable flexibility, the approach depends on the quality and breadth of available functions, which may limit its applicability in highly specialized domains.

Future Directions

The researchers highlighted several areas for future work, including:

1. Accuracy Management: Automatically distributing accuracy requirements across program components to optimize performance on NISQ devices.
2. Co-Design with Error Correction: Integrating logical quantum program design with error-correcting codes to enhance fault-tolerant computing.
3. Hybrid Quantum-Classical Programs: Extending their synthesis engine to handle mixed quantum and classical algorithms seamlessly.

The Road Ahead

As quantum computing moves closer to industrial applications, the need for robust, efficient software tools will grow. The EDA-inspired approach proposed by Classiq could serve as a blueprint for future developments in the quantum software stack, providing a pathway to practical and scalable solutions.

With further refinement, the researchers write that this technology could shift the focus of quantum software development from hardware constraints to functional design, unleashing the potential of quantum computing for broader applications.

For a more in-depth look at Classiq’s approach, the paper on arXiv can provide a much more technical look at the work. It’s important to remember that pre-print servers have not been officially peer-reviewed, but serve as a way for scientists to get informal feedback quickly, particularly in the fast-moving quantum industry.

Classiq’s team on this paper include: Tomer Goldfriend, Israel Reichental, Amir Naveh, Lior Gazit, Nadav Yoran, Ravid Alon, Shmuel Ur, Shahak Lahav, Eyal Cornfeld, Avi Elazari, Peleg Emanuel, Dor Harpaz, Tal Michaeli, Nati Erez, Lior Preminger, Roman Shapira, Erik Michael Garcell, Or Samimi, Sara Kisch, Gil Hallel, Gilad Kishony, Vincent van Wingerden, Nathaniel A. Rosenbloom, Ori Opher, Matan Vax, Ariel Smoler, Tamuz Danzig, Eden Schirman, Guy Sella, Ron Cohen, Roi Garfunkel, Tali Cohn, Hanan Rosemarin, Ron Hass, Klem Jankiewicz, Karam Gharra, Ori Roth, Barak Azar, Shahaf Asban, Natalia Linkov, Dror Segman, Ohad Sahar, Niv Davidson, Nir Minerbi, and Yehuda Naveh.