Multi-agent Learning in Heterogeneous Linear Contextual Bandit

This repos implements the H-LinUCB algorithm, publish in

Multi-Agent Learning with Heterogeneous Linear Contextual Bandits. (NeurIPS 2023)
Anh Do, Thanh Nguyen-Tang, Raman Arora

Training

• Use ./run.sh to generate the temporary results, we store them in results folder. We ran the script on a cluster with 30 cores CPU.
• Use plot.ipynb to plot the figure.

Simulation results