On Sampling and Inference using Quantum Algorithms

TL;DR

This paper explores the potential of quantum algorithms like Quantum Annealing and QAOA for sampling and inference in Markov networks, supported by extensive simulation results.

Contribution

It provides an overview of quantum sampling methods and presents simulation-based insights into their effectiveness for probabilistic inference.

Findings

Quantum algorithms show promise for Gibbs sampling.

Simulations indicate potential advantages over classical methods.

Results highlight challenges and future directions in quantum sampling.

Abstract

Quantum computers are projected to handle the Gibbs sampling and the related inference on Markov networks effectively. Apart from noting the background information useful for those starting the explorations in this important thread of Quantum Machine Learning, we capture some results and observations obtained through extensive simulations with two popular paradigms of sampling based on Quantum Annealing and Quantum Approximate Optimization Algorithm.