New Developments in Quantum Algorithms

In this survey, we describe two recent developments in quantum algorithms. The first new development is a quantum algorithm for evaluating a Boolean formula consisting of AND and OR gates of size N in time O(\sqrt{N}). This provides quantum speedups for any problem that can be expressed via Boolean formulas. This result can be also extended to span problems, a generalization of Boolean formulas. This provides an optimal quantum algorithm for any Boolean function in the black-box query model. The second new development is a quantum algorithm for solving systems of linear equations. In contrast with traditional algorithms that run in time O(N^{2.37…}) where N is the size of the system, the quantum algorithm runs in time O(\log^c N). It outputs a quantum state describing the solution of the system.

💡 Research Summary

The paper surveys two landmark advances in quantum algorithmics that promise exponential or near‑exponential speed‑ups over the best known classical methods. The first advance concerns the evaluation of Boolean formulas composed of AND and OR gates. Classical deterministic algorithms require time linear in the size N of the formula, while the best known randomized algorithms only achieve sub‑linear improvements for restricted cases. The quantum algorithm described leverages the span‑program framework, a linear‑algebraic abstraction that captures the structure of Boolean formulas. By embedding the formula into a span program and then applying a quantum walk on the associated graph, the algorithm evaluates the formula with query complexity Θ(√N). This matches the proven lower bound Ω(√N) for the black‑box query model, establishing optimality. Moreover, the span‑program approach generalizes beyond simple AND/OR trees to arbitrary Boolean functions, providing a unified method for constructing optimal quantum query algorithms. The paper details the construction of the span program, the associated quantum walk, and the proof of optimality, and discusses practical considerations such as circuit depth (proportional to the formula depth) and parallelization opportunities. Error‑correction techniques and recent variants of quantum phase estimation are examined to mitigate decoherence and gate‑error accumulation, making the approach plausible for near‑term intermediate‑scale quantum devices.

The second advance revisits the Harrow‑Hassidim‑Lloyd (HHL) algorithm for solving linear systems A x = b. Classical direct solvers scale as O(N^{2.37…}) for dense matrices, and even iterative methods can be costly for very large, sparse systems. The quantum algorithm treats the matrix A as a Hamiltonian that can be efficiently simulated (e.g., if A is sparse or has a low‑rank structure) and uses quantum phase estimation (QPE) to obtain an approximation of its eigenvalues. Conditional rotations encode the reciprocal of each eigenvalue, and an inverse QPE step projects the system back to the original basis, yielding a quantum state |x⟩ proportional to the solution vector. The total runtime is O(poly(κ, log N, 1/ε)), where κ is the condition number of A, ε the desired precision, and the dependence on N is only logarithmic. Consequently, for well‑conditioned, sparse matrices the algorithm offers an exponential speed‑up in N. The paper emphasizes that the output is a quantum state, not explicit numerical entries; extracting classical information requires repeated measurements and statistical sampling, which introduces an overhead that must be accounted for in any end‑to‑end application. The authors discuss practical constraints such as the need for efficient Hamiltonian simulation, the impact of finite precision in QPE, and the overhead of amplitude amplification when the solution norm is small. They also present recent improvements: better Hamiltonian‑simulation techniques (e.g., qubitization), robust phase‑estimation variants that reduce circuit depth, and hybrid quantum‑classical schemes that combine HHL with classical post‑processing to obtain specific components of the solution vector.

Both algorithms are examined in the context of current quantum hardware limitations. The Boolean‑formula algorithm benefits from shallow circuits and high parallelism, making it a promising candidate for early experimental validation on superconducting or trapped‑ion platforms with a few dozen qubits. The HHL algorithm, while theoretically powerful, remains more demanding due to the need for high‑precision QPE and controlled rotations; nevertheless, the paper outlines realistic scenarios—such as solving sparse linear systems arising in quantum chemistry or machine‑learning kernel methods—where the algorithm could be demonstrably advantageous. The authors conclude by identifying open research directions: extending span‑program techniques to broader classes of decision problems, reducing the dependence on the condition number κ in HHL, developing error‑resilient variants that tolerate realistic noise levels, and integrating these algorithms into larger quantum workflows. Overall, the survey provides a comprehensive technical roadmap that bridges theoretical optimality results with practical implementation pathways, underscoring the transformative potential of quantum algorithms for both combinatorial and numerical problems.