Efficient Graph Bandit Learning with Side-Observations and Switching Constraints

Authors: Xueping Gong, Jiheng Zhang

AAAI 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive numerical experiments on various types of graphs, including two real-world datasets, demonstrate the efficacy of our proposed methods and their advantages over benchmark methods in graph bandit settings.
Researcher Affiliation Academia Xueping Gong, Jiheng Zhang The Hong Kong University of Science and Technology, Clear Water Bay, Hong Kong S.A.R., China. EMAIL, EMAIL
Pseudocode Yes Algorithm 1: UCB-GGmax Algorithm 2: LP-GG
Open Source Code No The paper does not contain any explicit statement or link indicating that the source code for the methodology described is publicly available.
Open Datasets Yes we conduct numerical experiments on the dataset of roads in a small area of California (Leskovec and Krevl 2014). We conduct numerical experiments on the dataset of products in Amazon (Leskovec and Krevl 2014) Leskovec, J.; and Krevl, A. 2014. SNAP Datasets: Stanford Large Network Dataset Collection. http://snap.stanford.edu/ data.
Dataset Splits No The paper describes how reward distributions are initialized and graphs are generated for simulations, and mentions the use of real-world datasets, but it does not specify any training/test/validation splits for these datasets.
Hardware Specification No The paper does not provide specific details about the hardware (e.g., GPU models, CPU types, memory) used to run the experiments. It only mentions running simulations.
Software Dependencies No The paper mentions implementing Q-learning and its hyperparameters but does not specify any software libraries, frameworks, or their version numbers used for the implementation or experiments.
Experiment Setup Yes Q-learning (Sutton and Barto 2018): we choose the learning rate 0.5, the discount factor 0.9 and the exploration probability min(1, 2|V |/T).