nick-jhlee/bandits-network

This repository provides the codes for reproducing the experiments described in the paper Flooding with Absorption: An Efficient Protocol for Heterogeneous Bandits over Complex Networks (OPODIS 2023, best student paper) by Junghyun Lee, Laura Schmid, and Se-Young Yun.

If you plan to use this repository or refer to our work, please use the following bibtex format:

@InProceedings{lee2024absorption,
  author =	{Lee, Junghyun and Schmid, Laura and Yun, Se-Young},
  title =	{{Flooding with Absorption: An Efficient Protocol for Heterogeneous Bandits over Complex Networks}},
  booktitle =	{27th International Conference on Principles of Distributed Systems (OPODIS 2023)},
  pages =	{20:1--20:25},
  series =	{Leibniz International Proceedings in Informatics (LIPIcs)},
  ISBN =	{978-3-95977-308-9},
  ISSN =	{1868-8969},
  year =	{2024},
  volume =	{286},
  editor =	{Bessani, Alysson and D\'{e}fago, Xavier and Nakamura, Junya and Wada, Koichi and Yamauchi, Yukiko},
  publisher =	{Schloss Dagstuhl -- Leibniz-Zentrum f{\"u}r Informatik},
  address =	{Dagstuhl, Germany},
  URL =		{https://drops.dagstuhl.de/entities/document/10.4230/LIPIcs.OPODIS.2023.20},
  URN =		{urn:nbn:de:0030-drops-195100},
  doi =		{10.4230/LIPIcs.OPODIS.2023.20},
  annote =	{Keywords: multi-armed bandits, multi-agent systems, collaborative learning, network protocol, flooding}
}

Reproducing Figures

Figure 3, 4, 5

Run

python main.py

to obtain all the results. This will create Figure 4.

Then follow through the colab notebook plotting.ipynb to obtain Figures 3 and 5.

Figure 6

Run

python main_deltas.py

Then, go to the deltas folder and run

python computing_deltas.py

and then follow through the colab notebook plot_delta_regretgap.ipynb