Source code for "Artificial Agents Facilitate Human Cooperation through Indirect Reciprocity"

Published as a main track paper for ECAI 2024 (https://www.ecai2024.eu/).

Available here.

This folder contains all the source code to reproduce the results of the paper.

Requirements

The code base is divided between Julia 1.10 and Mathematica 14.0.

The Julia packages can be added via the command

] add Plots CairoMakie Colors Measures LaTeXStrings Memoization ArnoldiMethod SpecialFunctions

The Julia code also makes use of the following standard libraries: Dates, SparseArrays

Code Structure

Most of the code files do the mathematical operations described in the paper, consisting of calculating the cooperation index for a given set of parameters. The difference is in the parameters being studied. The structure of the code is as follows:

• ir_dynamics_with_AAs.nb - This Mathematica notebook is used to plot the simplexes with the reputation, stationary distribution and cooperation index at each state (which are then edited to form Fig 3, 4 of the main text);
• cooperation_study.jl - This julia file calculates the cooperation index for all norms as a function of the fraction of AAs, storing their values in text format in a folder. It also generates a simplified plot of the cooperation index (Fig 2, 5 of the main text);
• b_study.jl - This is used to study the cooperation index as a function of b, the cooperation benefit, for each norm. It produces the plot seen in Figure 1 of the Supplementary Material;
• grad_disc_alld_study.jl - This is used to study the gradient of selection along the AllD-Disc edge, for each norm. It produces the plots seen in Figure 5 and 7 of the Supplementary Material;
• error_study.jl - It contains the code to study the impact of different errors across social norms. It produces the results for the plots in Figures 10-13 of the Supplementary Material;
• reputation_dynamics.jl - Contains all the math to calculate reputation dynamics given a strategy state;
• strategy_dynamics.jl - Contains all the math to calculate strategy dynamics and the resulting cooperation/reputation distributions;
• utils.jl - Contains utility code used throughout the code base, pertaining to data processing;

Running the Julia code produces a text file in the "Plots/" directory, which is then read to make the plots.

Additionally, code is provided to replicate the plotting style used in the article. These are: coop_plots.jl, coop_plots_makie.jl, error_plots.jl. These read the "Plots/" directory to generate the plot