Robust Stochastic Bandit Algorithms under Probabilistic Unbounded Adversarial Attack

Authors: Ziwei Guan, Kaiyi Ji, Donald J. Bucci Jr., Timothy Y. Hu, Joseph Palombo, Michael Liston, Yingbin Liang4036-4043

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

Reproducibility Variable Result LLM Response
Research Type Experimental We provide multiple synthetic simulations of the proposed algorithms to verify these claims and showcase the inability of existing techniques to achieve sublinear regret. We also provide experimental results of the algorithm operating in a cognitive radio setting using multiple software-defined radios.
Researcher Affiliation Collaboration 1The Ohio State University, ECE Department 2015 Neil Ave, Columbus, OH 43210 2Lockheed Martin Advanced Technology Laboratories Cherry Hill, NJ, 08002, USA
Pseudocode Yes Algorithm 1 med-E-UCB ... Algorithm 2 med-ϵ-greedy
Open Source Code No The paper does not provide a link to or explicitly state the availability of open-source code for the described methodology.
Open Datasets No In our experiment, we choose the number of arms to be 10. The reward distribution of the ith arm is N(2i, 1) for i [K]... The rewards of the 5 channels without adversarial perturbation are normally distributed with mean SINR of [41, 37, 35, 31, 28] d B and unit variance.
Dataset Splits No The paper describes the generation of synthetic data or the setup for a live cognitive radio testbed, but does not specify train/validation/test dataset splits needed for reproducibility in the context of a fixed dataset.
Hardware Specification Yes Each radio node uses an Ettus Research Universal Software Radio Peripheral (USRP) B200 Software Defined Radio (SDR).
Software Dependencies No The paper mentions 'software-defined radio' but does not specify any particular software libraries, frameworks, or their version numbers used in the experiments.
Experiment Setup Yes For med-E-UCB, we set b = 4, ω = 4, and G = 103. For med-ϵ-greedy, we set c = 10. Other parameters are set as suggested by the original references.