Related Experiment Video
Updated: Feb 16, 2026
05:54
Eye-tracking to Distinguish Comprehension-based and Oculomotor-based Regressive Eye Movements During Reading
Published on: October 18, 2018
6.8K
Building an ACT-R Reader for Eye-Tracking Corpus Data
1Institute for Logic, Language and Computation, University of Amsterdam.
Topics in Cognitive Science
|December 19, 2017
Summary
This study introduces an ACT-R reader model capable of analyzing large eye-tracking datasets. The novel Bayesian estimation approach provides a better fit to cognitive process data than previous methods.
Area of Science:
- Cognitive Science
- Computational Neuroscience
- Psychology
Background:
- Cognitive architectures are typically applied to limited experimental data.
- Modeling large-scale eye-tracking corpus data presents unique challenges.
- Previous models often relied on manual parameter tuning.
Purpose of the Study:
- To develop an ACT-R reader model for large-scale eye-tracking corpus data.
- To improve the accuracy of cognitive models by utilizing extensive datasets.
- To establish a more robust parameter estimation method for cognitive architectures.
Main Methods:
- Development of a specialized ACT-R reader.
- Application to a large eye-tracking corpus dataset.
- Utilizing Bayesian estimation and Markov-Chain Monte Carlo (MCMC) for parameter estimation.
Main Results:
- The ACT-R reader model demonstrated a good fit to low-level processing data.
- Bayesian estimation and MCMC provided a more objective parameter estimation.
- The model's performance surpassed previous approaches in fitting corpus data.
Conclusions:
- The developed ACT-R reader model effectively models large-scale eye-tracking data.
- The Bayesian estimation method offers a superior alternative to manual parameter selection.
- This methodology is generalizable to other ACT-R models and datasets.
