Quantum Factorisation: Simpler Binomial Products

Pawel Wocjan and colleagues at theWatson Research Center, investigate a generalisation of classical multinomial coefficients, introducing a ‘twisted’ version weighted by pair-dependent values determined by a skew-symmetric matrix. Under specific structural conditions on this matrix, termed predecessor-uniformity, these twisted multinomial coefficients factorise into a product of Gaussian binomials with independent parameters. This extends existing formulas and represents a purely combinatorial identity applicable across the complex plane. Driven by applications in quantum computation, specifically pilot state preparation within Hamiltonian Decoded Quantum Interferometry (HDQI), this factorisation yields an exact matrix product state with a defined bond dimension, representing a key step towards developing algorithms for preparing complex quantum states.

Reduced computational complexity via factorisation of twisted multinomial coefficients

Computational cost for preparing pilot states in Hamiltonian Decoded Quantum Interferometry (HDQI) has been reduced from exponential scaling with the size of the largest connected component in an anticommutation graph, to a fixed bond dimension of just k+1. This represents a key advance, enabling exact matrix product state representation where previously it was impossible, as the prior method’s computational demands grew rapidly with problem size. This breakthrough stems from a new factorisation identity for ‘twisted’ multinomial coefficients, extending the standard formula for q-multinomial coefficients to accommodate m-1 independent parameters instead of one. The factorisation is purely combinatorial, holding true for any non-zero complex numbers without algebraic constraints, broadening its potential applications.

Extending the standard formula, a factorisation identity for ‘twisted’ multinomial coefficients now incorporates m-1 independent parameters instead of one. Achieving a fixed bond dimension of k+1, this advancement allows for exact matrix product state (MPS) representation, a sharp improvement over previous methods limited by exponential scaling with problem size. Functioning with any non-zero complex numbers without algebraic restrictions, the newly discovered identity is purely combinatorial, broadening its applicability beyond the initial quantum computing context. A greedy algorithm was also developed to determine predecessor-uniformity, a structural condition enabling this factorisation, in just O(m3) time. However, this currently addresses only pilot state preparation within the broader Hamiltonian Decoded Quantum Interferometry pipeline; efficient decoding of the associated Hamiltonian code remains a vital, unresolved challenge for practical implementation.

Pawel Wocjan and colleagues have demonstrated a new factorisation identity for ‘twisted’ multinomial coefficients. This identity allows for exact matrix product state representation, overcoming limitations of previous methods. The factorisation is purely combinatorial and applicable to any non-zero complex numbers. A greedy algorithm efficiently determines the necessary structural condition, predecessor-uniformity, in O(m3) time. However, further research is needed to address the efficient decoding of the Hamiltonian code within the broader HDQI framework.

Efficient quantum state preparation via factorised matrix product states

Matrix product states, or MPS, provide a compact way of representing complex quantum states, analogous to compressing a large image file without significant loss of detail; these form the core of this advance. The mathematical structure of these MPS was used to represent the coefficients needed for quantum state preparation, dramatically reducing computational demands. Identifying conditions, specifically ‘predecessor-uniformity’ within a ‘skew-symmetric matrix’, a grid of numbers where swapping rows and columns changes the sign of each entry, allowed for a key factorisation.

This factorisation breaks down a complex calculation into a product of simpler components, enabling an efficient matrix product state representation with a defined, manageable size. Developing a method utilising matrix product states, or MPS, allowed for efficient representation of quantum states and reduction of computational demands. This approach hinges on the concept of ‘predecessor-uniformity’ within a ‘skew-symmetric matrix’, allowing for factorisation of complex calculations. The resulting matrix product state has a bond dimension of k+1, representing the expansion coefficients of a twisted algebra, contrasting with existing methods which scale exponentially with the size of the largest component in an anticommutation graph.

Factorisation properties refine pilot state preparation in Hamiltonian Decoded Quantum

A new mathematical tool extending the well-known $q$-multinomial coefficient offers a potentially more efficient way to represent the complex quantum states needed for advanced algorithms. Investigations into Hamiltonian Decoded Quantum Interferometry (HDQI) reveal a surprising factorization property under specific conditions termed ‘predecessor-uniformity’, stemming from this work. Efficiently decoding the quantum information within the algorithm’s Hamiltonian code remains a vital, open question, however, as the current findings only address the initial stage of HDQI, pilot state preparation.

Despite addressing only the initial, pilot stage of a complex quantum process, this mathematical advance does not diminish its significance. This new tool arose while studying pilot state preparation in Hamiltonian Decoded Quantum Interferometry (HDQI), a recently proposed quantum algorithm for preparing Gibbs and ground states. The algorithm requires expanding a Hamiltonian to a certain power, and the coefficients, the pilot state amplitudes, are sums of twisted multinomial coefficients weighted by products of the Hamiltonian’s coefficients.

Under a structural condition on a skew-symmetric matrix Ω, predecessor-uniformity, the twisted multinomial factorizes as a product of Gaussian binomials with site-dependent parameters. This factorization extends the standard formula for the q-multinomial coefficient to multiple independent parameters, yielding an exact matrix product state of bond dimension k+1 for the expansion coefficients. However, efficient decoding of the associated Hamiltonian code is also required for physical Hamiltonians, and identifying such Hamiltonians remains an open problem.

This new ‘twisted multinomial coefficient’ exhibits a surprising factorization property under specific conditions, simplifying complex calculations. This mathematical development provides a new way to calculate coefficients important for preparing quantum states within Hamiltonian Decoded Quantum Interferometry, or HDQI, a technique aiming to create complex states for advanced computation. By generalising a known mathematical tool, the multinomial coefficient, a surprising factorization property was uncovered when specific conditions, termed predecessor-uniformity, are met within a weighting matrix. This factorization simplifies calculations, allowing for a compact representation of quantum states using a technique called a matrix product state, or MPS, which reduces computational demands.

The researchers discovered that a newly defined ‘twisted multinomial coefficient’ simplifies under specific conditions, allowing it to be expressed as a product of Gaussian binomials. This mathematical result provides a more efficient method for calculating coefficients used in pilot state preparation for Hamiltonian Decoded Quantum Interferometry, a quantum algorithm for creating complex states. The factorization also yields an exact matrix product state of bond dimension k+1, offering a compact way to represent these quantum states. The authors focused on this initial step of the HDQI process, leaving further development of the full protocol for future work.