PsiQuantum’s quantum circuits aid porous material simulation

PsiQuantum researchers have devised a method to represent the complex geometry of materials within a quantum computer, generating a realistic material structure on a 512x512x32 grid. This advance addresses a critical bottleneck in materials science; simulating materials with numerous microscopic features is often too computationally expensive for conventional computers, hindering progress in fields like fluid dynamics and mechanical engineering.

The team, led by Alice Barthe, created a “Quantum Cheese” algorithm that procedurally generates these features, efficiently loading complex geometries using reversible circuits rather than storing them. “The number of pores grows with the volume,” Barthe explains, highlighting the need for a scalable approach to unlock quantum advantage in solving material-related problems. The work is published as a preprint on arXiv.

Beyond classical limits of material simulation

Construct, PsiQuantum’s integrated software platform, now incorporates a method for generating the geometric descriptions needed to simulate materials on a quantum computer, even when those materials contain a vast number of microscopic features. The constructions are implemented and simulated using PsiQuantum Construct, connecting the mathematical description to reversible circuits and resource estimates. A companion notebook is available to follow the process of constructing these circuits.

This efficient geometry representation is particularly important for quantum algorithms designed to solve partial differential equations, a common technique in materials science. Classical methods for solving these equations typically scale linearly with the number of grid points used to represent the material, limiting the achievable resolution. Quantum algorithms, under certain conditions, can achieve sublinear scaling, potentially allowing for simulations with significantly higher resolution.

However, this speed-up is contingent on having an efficient way to input the material’s geometry into the quantum computer; a challenge the Quantum Cheese algorithm directly addresses. The team’s work focuses on simulating a larger volume of material at a fixed microscopic resolution, rather than simply increasing the resolution of a smaller volume. The ability to efficiently access different material realizations is achieved by promoting the classical key used in the pseudorandom permutation to a quantum register.

This allows algorithms to investigate distributions of geometries, opening up new avenues for studying material properties and behaviour. Alice Barthe notes that a complete application requires additional steps. “The next step is to bring these pieces together for concrete physical problems, e.g., to study the mechanical properties of porous materials, or the flow of fluid through them.” The researchers emphasize that their approach is not limited to materials with a fixed local structure.

The number of microstructures and total sites can grow, while the underlying local rule remains constant, effectively increasing the material volume at a fixed microscopic resolution, PsiQuantum says. This is demonstrated by a family of materials defined in the paper, where geometry oracles on exponentially increasing volumes can be implemented with polynomial-size.

The four steps of quantum cheese making involve associating each coordinate with a unique address, shuffling the addresses, choosing an integer M and flagging the coordinates that have addresses lower than this. The main challenge is to be able to describe exponentially many holes without ever explicitly listing all of them.

They demonstrated worst-case query lower bounds for amplitude-encoded formation fields and databases of holes, highlighting the importance of the specific approach taken with the Quantum Cheese algorithm. “For a fixed kernel with bounded integer coefficients, the oracle cost is polynomial in the number of qubits, even when the number of microfeatures is exponential,” the researchers report, emphasising the algorithm’s efficiency. The work represents a specific component of quantum simulation, but lays the groundwork for more complex applications in the future.

Quantum Cheese generates complex geometries

The “Quantum Cheese” algorithm, developed by PsiQuantum researchers, allows for the efficient procedural generation of complex material geometries within a quantum computing framework. Rather than explicitly storing the coordinates of microscopic features like pores, the algorithm constructs a geometry oracle that creates these features on demand, significantly reducing the computational resources needed for simulation. This approach addresses a core challenge in quantum materials science; the exponential growth in data required to represent increasingly detailed material structures.

The algorithm’s efficiency stems from its ability to define materials through local rules, rather than exhaustive lists of coordinates. A small set of parameters, termed a “kernel”, dictates how features are generated across a grid. This kernel, combined with a pseudorandom process, creates a textured geometry without requiring exponential storage, and the team demonstrated this by generating geometries on a 512x512x32 grid, effectively reproducing realistic material structures.

This means that as the complexity of the simulated material increases, the computational cost of accessing its geometry grows at a manageable rate, according to PsiQuantum. The procedural generation of geometry isn’t limited to simple features like holes; the framework extends to more complex textures and patterns. By defining a “texture” rather than a simple binary presence or absence of a feature, the algorithm can create materials with multiple labels and complex internal structures.

This versatility is important for simulating a wider range of materials, from porous media to complex foams. The researchers visualized this process, showing how a kernel, a pseudorandom binary field, and a convolution process combine to create a greyscale image representing the material’s geometry. The team’s work provides a concrete example of how to build and cost an efficient quantum circuit for accessing material geometry.

They created several geometries for which they could explicitly compile a polynomially-sized quantum geometry oracle, allowing for detailed inspection and analysis of the computational cost. This explicit construction is a starting point for developing more complex quantum simulations of material properties.

The ability to efficiently access geometric information is particularly important for quantum partial-differential-equation solvers, which could unlock quantum advantages in solving material-related problems like fluid flow through porous media or the mechanical properties of foams. The development of this efficient geometry oracle is a key step toward realizing the full potential of quantum computing in materials discovery and design.

Existing methods struggle with detail

Representing the detailed internal structure of materials for quantum simulation has historically relied on methods that quickly become impractical as complexity increases. Classical approaches, while effective for simpler geometries, demand exponentially more computational resources when tasked with defining materials containing a vast number of microscopic features, limiting the size and realism of simulations.

The challenge isn’t simply one of computational power, but of efficiently encoding geometric information into a form usable by quantum algorithms. Many quantum algorithms for solving partial differential equations, essential for modeling material behavior, assume access to a “geometry oracle”, a circuit that can coherently determine whether a given point lies within the material or in a void.

PsiQuantum researchers addressed this by focusing on procedural generation of these features, rather than explicit storage. Their solution designs a geometry oracle that procedurally generates exponentially many microscopic features. The core of the “Quantum Cheese” algorithm involves converting each coordinate into a unique integer address, then shuffling those addresses using a pseudorandom reversible permutation. This permutation, built using a lightweight cipher, introduces the necessary randomness to create realistic pore structures.

By comparing the shuffled address to a chosen number, M, the algorithm determines whether a coordinate should be the center of a pore. The spatial arrangement of these pores is then defined by a fixed “hole shape,” or stencil, applied around each center. This method’s efficiency hinges on locality. The algorithm only needs to examine the immediate neighborhood around a given coordinate to determine its status, avoiding the need to search a global list of pores. The resulting coordinates constitute the centers of the holes.

Simulations need further validation

Construct, PsiQuantum’s software platform, facilitated the implementation and simulation of these geometry-generating constructions, allowing for a direct link between mathematical descriptions and reversible circuits with associated resource estimates. This ability to move beyond simple examples is important, as the true power of the technique lies in its potential to model complex, realistic materials. Beyond the immediate concerns of circuit fidelity, the method’s versatility also introduces new challenges for validation.

The current implementation relies on a keyed construction based on a lightweight cipher for generating pseudorandomness, a critical component for shuffling coordinate addresses and introducing spatial disorder. Any biases or patterns in the pseudorandom number generation could potentially propagate through the simulation, leading to inaccurate results. However, efficiently storing and accessing this information within a quantum computer remains a challenge. This is particularly important for simulating large volumes of material, where the amount of data can quickly become prohibitive.

The ability to efficiently generate exponentially many microscopic features is a key advantage of this approach, but it also raises questions about the limits of procedural generation. However, realizing that potential requires continued validation and refinement of the underlying algorithms and hardware.

Unlocking quantum advantage for materials

Alice Barthe of PsiQuantum has designed a method to efficiently represent the complex geometry of materials, a key step toward simulating their behavior on quantum computers. This advance focuses on efficiently supplying structured geometric information, a specific component of the broader challenge of quantum simulation. Rather than explicitly listing the coordinates of every feature, the algorithm procedurally generates them, significantly reducing the computational burden. The framework defines textures more generally than a simple list of offsets, allowing for more complex geometries.

This is achieved through a convolution with a local kernel, shaping the texture and enabling the creation of diverse microstructures. The research addresses a central question in quantum materials science: when can geometry oracles be implemented by polynomially sized circuits?

The team’s work provides a positive answer for a specific family of materials with exponentially many microscopic features. The next phase of research will focus on integrating these geometric representations into complete quantum simulations. A complete application requires not only efficient geometry access but also the construction of the physical operator, the solution of relevant equations and the extraction of useful quantities.

More like this

Quantum AlgorithmsData alone could fix quantum circuit design choices before training

Quantum AlgorithmsQuantum algorithm cuts cloud droplet calculation costs, study shows

Quantum Research NewsOne clean qubit can control three qubits and stay separable

Quantum PhysicsQueen’s University releases open-source tool for quantum waveguide simulations