Learning rapid and precise skills.

Area of Science:

• Cognitive Science
• Computational Psychology
• Human-Computer Interaction

Background:

• Human skill acquisition is complex, involving both instruction and practice.
• Existing models may not fully capture the interplay between declarative and procedural knowledge.
• Understanding the cognitive mechanisms behind skill learning is crucial for effective training.

Purpose of the Study:

• To present a unified theory of skill acquisition based on the ACT-R architecture.
• To develop and validate a computational model simulating human learning of complex skills.
• To investigate the role of specific architectural components, including a novel Controller module, in skill acquisition and transfer.

Main Methods:

• Development of a computational model based on the ACT-R architecture, incorporating motor timing, instruction interpretation, knowledge compilation, and a new Controller module.
• Experimental validation using two games derived from the video game Space Fortress.
• Testing predictions of the Controller module regarding transfer of learning between video games.

Main Results:

• The computational model successfully simulated human learning across three video games with consistent parameter settings.
• Experimental results confirmed the model's predictions regarding skill acquisition and the lack of transfer.
• The four key aspects of the ACT-R architecture (motor timing, declarative interpretation, procedural compilation, and the Controller module) were shown to be critical.

Conclusions:

• The proposed theory and its computational implementation provide a robust framework for understanding human fluency in complex skills.
• The ACT-R architecture, particularly with the addition of the Controller module, effectively models the integration of instruction and experience in learning.
• The findings have implications for designing more effective training programs and understanding cognitive skill development.