Randomised Gaussian Process Upper Confidence Bound for Bayesian Optimisation

Authors: Julian Berk, Sunil Gupta, Santu Rana, Svetha Venkatesh

IJCAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental In this section we present results that demonstrate the performance of RGP-UCB in comparison to other common acquisition functions. We also demonstrate the impact of varying the θ parameter of the gamma distribution used to sample βt. The Python code used for this paper can be found at https://github.com/jmaberk/RGPUCB. We test our method against a selection of common acquisition functions on a range of Bayesian optimisations problems. These include a range of synthetic benchmark functions and real-world optimisation problems. These are all transformed into continuous maximisation problems for consistency.
Researcher Affiliation Academia Julian Berk , Sunil Gupta , Santu Rana and Svetha Venkatesh Applied Artificial Intelligence Institute {jmberk, sunil.gupta, santu.rana, svetha.venkatesh}@deakin.edu.au
Pseudocode Yes Algorithm 1 Bayesian Optimisation with RGP-UCB
Open Source Code Yes The Python code used for this paper can be found at https://github.com/jmaberk/RGPUCB.
Open Datasets Yes All benchmark functions use the recommended parameters from https://www.sfu.ca/ ssurjano/optimization.html and All experiments are done with the public Space GA scale dataset 2. Dataset can be found at https://www.csie.ntu.edu.tw cjlin/ libsvmtools/datasets/regression.html
Dataset Splits No The paper does not provide explicit training/validation/test dataset splits (e.g., percentages or sample counts).
Hardware Specification No The paper does not provide any specific hardware details (e.g., GPU/CPU models, memory) used for running the experiments.
Software Dependencies No The paper mentions 'Python code' but does not specify any software dependencies with version numbers.
Experiment Setup Yes In each case, the experiment was run for 40d iterations and repeated 10 times with 3d+1 different initial points. The initial points are chosen randomly with a Latin hypercube sample scheme [Jones, 2001]. and We also demonstrate the impact of varying the θ parameter of the gamma distribution used to sample βt.