A Compass on the Quantum State Sphere: The Hopf Ansatz for Arbitrary Pure-State Optimization

The authors introduce the Hopf ansatz, a binary-tree circuit designed to represent arbitrary normalized real and complex quantum state vectors. By mapping internal tree angles to probability splits and leaf phases to complex degrees of freedom, the ansatz serves as a structured chart over the unit sphere in Hilbert space. This construction is bidirectional: it provides an explicit inverse map from amplitudes to physical angles and organizes the circuit skeleton to facilitate both state preparation and gradient access.

A primary contribution of this work is the realization of a diagonal pullback metric within the Hopf chart. In standard variational quantum algorithms, computing the natural gradient often requires estimating and inverting a dense quantum geometric tensor, which is computationally expensive. Because the Hopf ansatz diagonalizes this metric, Riemannian descent can be performed through simple, component-wise operations. The metric entries are determined by the probability mass entering specific tree nodes, providing a clear geometric interpretation of the search space.

The paper demonstrates that coordinate tangents—the directions in which the state changes with respect to parameters—can be prepared exactly as normalized quantum states using the same circuit skeleton. This allows for efficient gradient estimation for Hamiltonian objectives, where each component is expressed as a transition moment between the current state and a tangent state. Because the gradient-access configurations are organized by tree layer, the number of distinct circuit families required grows only logarithmically with the Hilbert-space dimension, offering a significant advantage for large-scale variational optimization.

This framework bridges the gap between universal state preparation and practical variational optimization. By turning the ansatz into a navigable coordinate system, the Hopf approach simplifies the implementation of natural gradient descent and improves the efficiency of variational quantum eigensolvers (VQE) and quantum metrology protocols. The ability to perform metric-aware updates without the overhead of full geometric tensor estimation makes it a powerful tool for high-precision quantum state optimization.