On the Finite-Time Performance of the Knowledge Gradient Algorithm

Authors: Yanwen Li, Siyang Gao

ICML 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Last, we use numerical experiments to compare the bounds we derive and the performance of the KG algorithm.
Researcher Affiliation Academia 1Department of Advanced Design and Systems Engineering, City University of Hong Kong, Hong Kong 2School of Data Science, City University of Hong Kong, Hong Kong.
Pseudocode Yes Algorithm 1 Knowledge Gradient
Open Source Code No The paper does not provide concrete access to source code, nor does it state that the code is available in supplementary materials or upon request.
Open Datasets No The paper defines 'Instance 1', 'Instance 2', etc. with specific parameters (µ, σ), which appear to be custom-generated for the experiments. It does not provide access information (link, DOI, citation) for these or any other dataset.
Dataset Splits No The paper does not provide specific dataset split information (exact percentages, sample counts, citations to predefined splits, or detailed splitting methodology) for training, validation, or testing data.
Hardware Specification No The paper does not provide specific hardware details (exact GPU/CPU models, processor types, or memory amounts) used for running its experiments.
Software Dependencies No The paper does not provide specific ancillary software details (e.g., library or solver names with version numbers) needed to replicate the experiment.
Experiment Setup Yes At the beginning, the algorithm pulls each arm for five times to obtain the initial estimates of the mean rewards of the arms. The numerical test is conducted on the following two instances. Instance 1. We consider a set of ten arms {1, 2, . . . , 10}. Set µi = 1 for i = 1, . . . , 9, µ10 = 2, and σi = 1 for i = 1, . . . , 10. The best arm b = 10. Instance 2. We consider a set of ten arms {1, 2, . . . , 10}. Set µi = 1 and σi = 1 for i = 1, . . . , 5, µi = 2 and σi = 2 for i = 6, . . . , 9, µ10 = 3 and σ10 = 3. The best arm b = 10.