Reinforcement_learning_workshop

This is an introduction to reinforcement learning of choice data. The material of this repository has been used for a workshop at the University of Granada on the 1st of June 2022. The workshop is divided into two parts: basic concepts and model fitting. The slides "reinforcement_learning_slides.pdf" constitutes a guide through the basic principles and applications of cognitive modeling in general and reinforcement learning models in particular. The .md files 1.basic_concepts.md and 2.model_fitting.md contain practical examples of the principles illustrated in the slides and can be viewed on github. The corresponding .Rmd files can be also run in Rstudio on a computer. Create.data.R contains a script that creates the syntetic data that can be found in the Data folder.

Main contents

• Introduction to cognitive modelling
• Introduction to reinforcement learning
• Pavlovian models
• Instrumental models
• Model simulation
• Parameter recovery
• Model recovery
• Model fitting and parameter estimation
• Model comparison
• Model Validation

Structure of the repository

• Data - data used to fit the models
• helper_functions - custom R functions used by the main scripts
• likelihood_functions - functions used to compute the likelihood of the choice data, given the models and the parameters
• model_recovery - custom scripts to do model recovery
• output_files - output of the scripts
• simulation_functions - functions used to simulate choice data given the models and the parameters