element-wise-transforms-quantum

name	element-wise-transforms-quantum
category	quantum
description	Quantum element-wise transforms methodology for efficient matrix operations — reduces space exponentially compared to prior work using polynomial function applied element-wise. Applications to ML, simulation, signal processing.
trigger	element-wise quantum transforms, quantum matrix operations, quantum element-wise, QSVT element-wise, polynomial matrix quantum
source	arXiv: 2606.06456
created	2026-06-09

Quantum Element-wise Transforms

Overview

This methodology constructs improved quantum algorithms for element-wise matrix transforms — applying polynomial functions element-wise to matrices embedded in block encodings. Space complexity reduced exponentially in polynomial degree compared to prior work.

Core Technique

Problem Setting

Given a matrix A block-encoded in a unitary U_A, apply polynomial function f element-wise to produce f(A).

Key Insight

• QSVT/LCU work on spectral transforms (eigenvalues)
• Element-wise transforms require position-register encoding and polynomial evaluation in computational basis
• Space: O(log k) qubits for degree-k polynomial vs O(k) prior (exponential improvement)

Algorithm Steps

1. Block encode input matrix A into unitary U_A
2. Prepare position registers for row and column indices
3. Apply controlled polynomial evaluation in computational basis
4. Uncompute auxiliary registers
5. Extract result via amplitude estimation or post-selection

Applications

• ML: Element-wise activation functions in QNNs, Hadamard products for kernels, attention components
• Simulation: Nonlinear Hamiltonian term transformations, perturbation corrections, density matrix manipulation
• Signal Processing: Element-wise windowing, nonlinear filtering, feature extraction

Pitfalls

• Element-wise ≠ spectral: QSVT applies to eigenvalues, not individual elements
• Block encoding overhead requires careful amplitude amplification
• Polynomial approximation quality affects output fidelity
• Post-selection probability may require amplitude amplification

Verification

• Test with small matrices for classical verification
• Check space complexity scaling with polynomial degree
• Verify on known element-wise transforms (ReLU, sigmoid approximations)