Blythe et al., "Massive Multi-agent Data-Driven Simulations of the GitHub Ecosystem" (2019)

2019-06-26 → 2026-08-14

Jim Blythe, John Bollenbacher, Di Huang, Pik-Mai Hui, Rachel Krohn, Diogo Pacheco, Goran Muric, Anna Sapienza, Alexey Tregubov, Yong-Yeol Ahn, Alessandro Flammini, Kristina Lerman, Filippo Menczer, Tim Weninger, and Emilio Ferrara, Advances in Practical Applications of Survivable Agents and Multi-Agent Systems: The PAAMS Collection, 3–15 (2019)
Link | PDF | Code

@inproceedings{blythe2019massive,
    author = {Jim Blythe and John Bollenbacher and Di Huang and Pik-Mai Hui and Rachel Krohn and Diogo Pacheco and Goran Muric and Anna Sapienza and Alexey Tregubov and Yong-Yeol Ahn and Alessandro Flammini and Kristina Lerman and Filippo Menczer and Tim Weninger and Emilio Ferrara},
    title = {Massive Multi-agent Data-Driven Simulations of the GitHub Ecosystem},
    booktitle = {Advances in Practical Applications of Survivable Agents and Multi-Agent Systems: The PAAMS Collection},
    volume = {11523},
    pages = {3--15},
    publisher = {Springer International Publishing},
    doi = {10.1007/978-3-030-24209-1_1},
    year = {2019},
}

Describes a data-driven, multi-agent simulation framework built for the DARPA SocialSim GitHub challenge. The system scaled to millions of users and repositories; models that sampled relatively stable user-specific action distributions performed broadly well across evaluation metrics.

Agent based model, GitHub, Computational social science

Receive my updates

YY's Random Walks — Science, academia, and occasional rabbit holes.

YY's Bike Shed — Sustainable mobility, urbanism, and the details that matter.

×