A Structural Model of Attentional Effort Dynamics: Evidence from a Naturalistic Discrimination Task

This repository contains code and data for the paper A Structural Model of Attentional Effort Dynamics: Evidence from a Naturalistic Discrimination Task. The repository is divided into two main sections:

1. Illustrative Example (synthetic data generation and parameter estimation)
2. Naturalistic Umpire Dataset (contextual inference and policy estimation for real umpires)

Repository Structure

1. Illustrative Example (illustrative_example/)

This folder contains synthetic data and scripts for generating and estimating model parameters.

• synthetic_data_observations.npy – Generated synthetic observations
• synthetic_data_outcomes.npy – Generated synthetic outcomes
• synthetic_data_generator.ipynb – Code to generate synthetic data
• synthetic_data_estimation.ipynb – Code to estimate model parameters from synthetic data (results referenced in the paper)

2. Naturalistic Umpire Dataset (naturalistic_umpire_dataset/)

This section contains real-world data for six umpires, structured as follows:

• Each umpire has a dedicated folder (umpire_1/, umpire_2/, ..., umpire_6/)
• Each umpire's folder contains:
  • context_inference_umpire_#.csv – Dataset for contextual inference (Phase 1)
  • utility_estimation_umpire_#.csv – Dataset for estimating attention and task policy (Phase 2, includes pre-trained info from Phase 1)

The main script for processing umpire data is:

• overall_estimation.ipynb – Code to reproduce umpire results

Usage Instructions

1. Synthetic Data Example

Run the following notebook in illustrative_example/:

• Estimate model parameters: synthetic_data_estimation.ipynb

2. Umpire Dataset Estimation

To reproduce results for umpires, navigate to naturalistic_umpire_dataset/ and run:

• overall_estimation.ipynb

Dependencies

Ensure you have the following Python libraries installed:

pip install numpy pandas matplotlib scipy jupyter