Adiabatic Quantum Kitchen Sinks for Learning Kernels Using Randomized Features

TL;DR

This paper introduces adiabatic quantum kitchen sinks, a novel hybrid quantum-classical algorithm that uses adiabatic quantum devices to transform data features non-linearly, improving binary classification performance on synthetic and real datasets.

Contribution

It proposes the adiabatic quantum kitchen sinks algorithm, leveraging adiabatic quantum devices for feature transformation in machine learning, a new approach compared to existing quantum algorithms.

Findings

Significantly improves classical linear classifier accuracy.

Demonstrates effectiveness on synthetic and real-world datasets.

Potential for implementation on current adiabatic quantum devices.

Abstract

Quantum information processing is likely to have far-reaching impact in the field of artificial intelligence. While the race to build an error-corrected quantum computer is ongoing, noisy, intermediate-scale quantum (NISQ) devices provide an immediate platform for exploring a possible quantum advantage through hybrid quantum--classical machine learning algorithms. One example of such a hybrid algorithm is "quantum kitchen sinks", which builds upon the classical algorithm known as "random kitchen sinks" to leverage a gate model quantum computer for machine learning applications. We propose an alternative algorithm called "adiabatic quantum kitchen sinks", which employs an adiabatic quantum device to transform data features into new features in a non-linear manner, which can then be employed by classical machine learning algorithms. We present the effectiveness of our algorithm for performing binary classification on both a synthetic dataset and a real-world dataset. In terms of classification accuracy, our algorithm significantly enhances the performance of a classical linear classifier on the studied binary classification tasks and can potentially be implemented on a current adiabatic quantum device to solve practical problems.