Anytime Exploration for Multi-armed Bandits using Confidence Information

Authors: Kwang-Sung Jun, Robert Nowak

ICML 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Our analysis shows that the sample complexity of AT-LUCB is competitive to anytime variants of existing algorithms. Moreover, our empirical evaluation on AT-LUCB shows that AT-LUCB performs as well as or better than state-of-the-art baseline methods for anytime Explore-m.
Researcher Affiliation Academia Kwang-Sung Jun KJUN@DISCOVERY.WISC.EDU Wisconsin Institutes for Discovery, UW-Madison, 330 N. Orchard St., Madison, WI 53715 USA Robert Nowak RDNOWAK@WISC.EDU Wisconsin Institutes for Discovery, UW-Madison, 330 N. Orchard St., Madison, WI 53715 USA
Pseudocode Yes Algorithm 1 AT-LUCB
Open Source Code No The paper does not provide an explicit link to open-source code for the methodology presented.
Open Datasets Yes We use the New Yorker dataset.6 The data consists of n = 496 captions with 100K ratings. Footnote 6: Dataset number 499 from https://github.com/nextml/NEXT-data/.
Dataset Splits No The paper describes the datasets used (toy MAB instances and New Yorker dataset) but does not specify training, validation, or test splits in detail for reproducibility.
Hardware Specification No The paper does not specify the hardware (e.g., GPU, CPU models) used for running the experiments.
Software Dependencies No The paper does not mention specific software dependencies with version numbers used for the experiments.
Experiment Setup Yes We run AT-LUCB with δ1 = 1/2, = .99, and = 0. We set the exploration parameter of UCB as 2. We run each method 200 times.