The U.S. Department of the Air Force (DAF) must develop and sustain mission-critical knowledge, skills, and abilities (KSAs) among airmen, guardians, and civilians. Yet training and education is costly. This report describes how computational cognitive models can deliver personalized and adaptive training that enhanced learning gains and reduced training time for mission-critical KSAs, with emphasis on second-language learning, for the DAF.

Exploring the Use of Computational Cognitive Models to Personalize Training

Matthew Walsh, Mark Toukan, Thomas Goode, Andrea M. Abler, Sean Mann, Lewis Schneider

ResearchPublished Jul 13, 2023

Cover: Exploring the Use of Computational Cognitive Models to Personalize Training
Download PDF
Additional Downloads

Includes Research Summary

The purpose of training and education in the United States Department of the Air Force (DAF) is to develop and sustain mission-critical knowledge, skills, and abilities (KSAs) among airmen, guardians, and civilians. The DAF must deliver effective training and education to fully use its human capital, provide warfighting assets to combatant commanders, and maintain asymmetric advantage over competitors. Yet training and education is costly. A recent budget request included more than $2 billion for training and education, and recent guidance has highlighted that the U.S. Air Force must transform all facets of training and education to field a highly capable force in an affordable manner.

This report focuses on computational cognitive models, a class of training technologies with transformative potential. Computational cognitive models emulate psychological processes like knowledge acquisition and retention. These models have been used to develop empirically grounded training curricula and deliver personalized training in diverse domains. The primary benefits of using these models to deliver personalized training are enhanced learning gains and reduced training time.

This report explores the feasibility of applying computational cognitive models to the acquisition and sustainment of mission-critical KSAs, with emphasis on second-language learning. The authors affirm that cognitive models can be integrated with training curricula in a variety of ways, and each of these potential courses of action (COAs) presents different levels of benefits along with different technical and logistical challenges.

Key Findings

  • Cognitive models are underexplored relative to other approaches to delivering adaptive training, yet they might offer significant benefits.
  • Furthermore, cognitive models provide a way to trace knowledge retention across long periods and during periods of disuse.
  • Of the computational cognitive models that have been proposed, the Predictive Performance Equation (PPE) has been demonstrated to be successful in many domains and settings.
  • Building on previous work that has tested PPE in real-world settings, PPE also provides a valid account of second-language learning in representative settings.
  • Of the elements covered in the Linguist Next (LN) Standard Arabic Basic course, PPE is most applicable to the acquisition and retention of task-critical vocabulary.
  • PPE can be used to deliver empirically grounded recommendations for when to introduce and rehearse task-critical vocabulary in the LN curriculum at the classroom level.
  • Student performance measures captured in the LN Standard Arabic Basic course provide information about mastery, yet they are not intended for personalized training.
  • One promising COA is to deliver rehearsal and assess mastery of task-critical vocabulary using a separate software application designed for cognitive model-based personalized training.

Recommendations

  • The DAF should leverage cognitively inspired technologies to augment training.
  • The DAF should use computational cognitive models of knowledge acquisition and retention to deliver empirically based, personalized training.
  • Air Force Research Laboratory (AFRL) should work with the DAF training enterprise to identify KSAs that underlie task performance to permit the application of cognitive models.
  • Furthermore, AFRL should work with the DAF training enterprise to tag training materials based on the KSAs they involve and to develop quantitative measures of KSA performance.
  • AFRL should continue to develop statistical methods to calibrate cognitive models efficiently and effectively for individual students and groups of students. Different implementations of the method used here, Bayesian Hierarchical models, are needed to scale methods to larger numbers of students.
  • AETC LN should use a separate software application specifically designed for personalized training to deliver rehearsal of task-critical vocabulary. Because learning events cannot be fully tailored for cognitive model-based, personalized training, a promising option would be to track all learning activities that students complete but to limit personalized training to a separate application that complements the full curriculum.

Topics

Document Details

Citation

Chicago Manual of Style

Walsh, Matthew, Mark Toukan, Thomas Goode, Andrea M. Abler, Sean Mann, and Lewis Schneider, Exploring the Use of Computational Cognitive Models to Personalize Training. Santa Monica, CA: RAND Corporation, 2023. https://www.rand.org/pubs/research_reports/RRA1565-1.html.
BibTeX RIS

Research conducted by

The research described in this report was commissioned by the Air Force Research Laboratory (AFRL) and conducted within the Workforce, Development, and Health Program of RAND Project AIR FORCE.

This publication is part of the RAND research report series. Research reports present research findings and objective analysis that address the challenges facing the public and private sectors. All RAND research reports undergo rigorous peer review to ensure high standards for research quality and objectivity.

This document and trademark(s) contained herein are protected by law. This representation of RAND intellectual property is provided for noncommercial use only. Unauthorized posting of this publication online is prohibited; linking directly to this product page is encouraged. Permission is required from RAND to reproduce, or reuse in another form, any of its research documents for commercial purposes. For information on reprint and reuse permissions, please visit www.rand.org/pubs/permissions.

RAND is a nonprofit institution that helps improve policy and decisionmaking through research and analysis. RAND's publications do not necessarily reflect the opinions of its research clients and sponsors.