rl_cheap_talk
Code for "D. Condorelli, M. Furlan (2024). Cheap Talking Algorithms"
Science Score: 54.0%
This score indicates how likely this project is to be science-related based on various indicators:
-
✓CITATION.cff file
Found CITATION.cff file -
✓codemeta.json file
Found codemeta.json file -
✓.zenodo.json file
Found .zenodo.json file -
○DOI references
-
✓Academic publication links
Links to: arxiv.org -
○Academic email domains
-
○Institutional organization owner
-
○JOSS paper metadata
-
○Scientific vocabulary similarity
Low similarity (13.8%) to scientific vocabulary
Repository
Code for "D. Condorelli, M. Furlan (2024). Cheap Talking Algorithms"
Basic Info
Statistics
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
- Releases: 0
Metadata Files
README.md
Overview
This repository contains the code for "D. Condorelli, M. Furlan (2024). Cheap Talking Algorithms". Replication instructions are listed in the replication file.
Table of Contents
Installation
Install Julia: Download and install Julia: julialang.org.
Install Dependencies: Run in the Julia REPL:
using Pkg Pkg.add("ArgParse") Pkg.add("TOML") Pkg.add("JLD2") Pkg.add("PrettyTables") Pkg.add("ProgressMeter") Pkg.add("Random") Pkg.add("StatsBase") Pkg.add("PGFPlotsX") Pkg.add("DataStructures")Clone the Repository: Clone this repository to your local machine.
bash git clone https://github.com/massimilianofurlan/rl_cheap_talk.git
Usage
Navigate to the Project Folder:
bash cd rl_cheap_talkBasic Usage: To run a batch of simulations use
bash julia --threads=auto --check-bounds=no main.jl [options]You can replace
autowith the desired number of threads. Options are primarily used to select the configuration of the game, such as the number of states, the extent of the sender's bias, the utility functions and the distribution over states. To get the complete list of options usebash julia main.jl --helpTo change the reinforcement learning hyperparameters, such as the learning rate and exploration rate, edit the config file. The section
[default]is loaded by default. A different section can be loaded using the--configoption.Others: To automatically run a batch of simulations for each bias level in {0.0, 0.01, ..., 0.49, 0.5} use
bash julia --threads=auto scripts/script1.jl --config=CONFIG_SECTION --out_dir=OUT_DIR --step_bias=0.01Replace
CONFIG_SECTIONwith the desired section from the config file andOUT_DIRwith the desired name for the output directory. Use--helpto get the complete list of options.After running the script, TikZ plots can be generated using
bash julia --threads=auto scripts/generate_plots_1.jl --in_dir=OUT_DIR --step_bias=0.01where
OUT_DIRmust match the output folder of the previous command. Also the increment in the bias must match che one used in the previous command. Generating plots requires an additional Julia dependency. To install it, run in the Julia REPLusing Pkg Pkg.add("PGFPlotsX")Note that PGFPlotsX also requires having a LaTeX installation on your machine.
License
This project is licensed under the GNU Affero General Public License v3.0 - see the LICENSE file for details.
For more information about the AGPL-3.0 license, please visit the GNU website.
Owner
- Name: Massimiliano Furlan
- Login: massimilianofurlan
- Kind: user
- Location: Coventry, United Kingdom
- Company: University of Warwick
- Website: https://massimilianofurlan.github.io
- Repositories: 5
- Profile: https://github.com/massimilianofurlan
MRes in Economics
Citation (CITATION.cff)
cff-version: 1.2.0 message: "If you use this software, please cite it as below." authors: - family-names: "Condorelli" given-names: "Daniele" - family-names: "Massimiliano" given-names: "Furlan" title: "Reinforcement learning in cheap-talk games." version: 1.0.0 date-released: 2023-09-25 url: "https://github.com/massimilianofurlan/rl_cheap_talk"