Quantum Bayesian Optimization

Authors: Zhongxiang Dai, Gregory Kang Ruey Lau, Arun Verma, YAO SHU, Bryan Kian Hsiang Low, Patrick Jaillet

NeurIPS 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We use simulations, as well as an experiment using a real quantum computer, to verify that the theoretical quantum speedup achieved by our Q-GP-UCB is also potentially relevant in practice.
Researcher Affiliation Academia 1Department of Computer Science, National University of Singapore 2CNRS@CREATE, 1 Create Way, #08-01 Create Tower, Singapore 138602 3Department of Electrical Engineering and Computer Science, MIT
Pseudocode Yes Algorithm 1 Q-GP-UCB
Open Source Code No The paper states: "We use the Qiskit python package to implement the QMC algorithm (Lemma 1) following the recent work of [20]." and discusses running experiments on IBM quantum computers, but does not provide a link to their own source code for Q-GP-UCB.
Open Datasets Yes In the Auto ML experiment, an SVM is used for diabetes diagnosis. That is, we adopt the diabetes diagnosis dataset which can be found at https://www.kaggle.com/uciml/ pima-indians-diabetes-database, which is under the CC0 license.
Dataset Splits Yes We use 70% of the dataset as the training set and the remaining dataset as the validation set.
Hardware Specification Yes All our experiments are run on a computer with Intel(R) Xeon(R) Gold 6326 CPU @ 2.90GHz, with 64 CPUs.
Software Dependencies Yes We use the Qiskit python package to implement the QMC algorithm (Lemma 1) following the recent work of [20].
Experiment Setup Yes For our Q-GP-UCB, we choose beta_s = 1 + log s, s >= 1 following the order given by the theoretical value of beta_s (Theorem 3). For classical GP-UCB, we tried different approaches to setting beta_s: beta_s = sqrt(2) and beta_s = 1, all of which are commonly adopted practices in BO; we found that beta_s = sqrt(2) and beta_s = 1 have led to the best performances for GP-UCB in, respectively, the synthetic and Auto ML experiments.