The authors reviewed academic and commercial literature to document how recommendation engines have been used to facilitate job matches. They then examined U.S. Air Force data stores and affordances of the Air Force Talent Marketplace to understand the types of recommendation engines they could support. Finally, they reviewed Air Force instructions to identify how recommendation engines could be incorporated into the assignment process.

Machine Learning-Enabled Recommendations for the Air Force Officer Assignment System

Volume 5

Avery Calkins, Monique Graham, Claude Messan Setodji, David Schulker, Matthew Walsh

ResearchPublished Mar 4, 2024

The Air Force Talent Marketplace provides a way for officers, position owners, and assignment teams to gain greater visibility into the assignment process, and to express and satisfy needs and preferences in a more transparent manner. While the Talent Marketplace has great potential, it may also introduce new challenges. For example, officers may not accurately gauge how different assignments may contribute to their development, position owners may need to vet long lists of candidates, and officer and position owner preferences may not meet the needs of the U.S. Air Force (USAF).

Recommendation engines are an established tool for presenting or ranking options from a large set based on a model of individual preferences or needs. These engines have been used in commercial job markets and might be used to enhance the Talent Marketplace.

The authors reviewed academic and commercial literature to document how recommendation engines have been used to facilitate job matches. They then examined USAF data stores and affordances of the Talent Marketplace to understand the types of recommendation engines they could support. Finally, they reviewed Air Force instructions to identify how recommendation engines could be incorporated into the assignment process.

Key Takeaways

  • With the advent of the Talent Marketplace, the USAF is well positioned to implement machine learning (ML)–enhanced assignment recommendations. Besides providing a platform to deliver recommendations, the Talent Marketplace can be used to gather additional information about officer qualifications, position details, and historic preferences, which would improve the effectiveness of future recommendations.
  • A useful way to implement ML-enhanced recommendations would be to organize them around the themes of officer development, performance, and job satisfaction because they capture the factors most likely to be important to officers, position owners, and the career field assignment team. Recommendations organized around these themes are also easier to construct and implement than a single holistic recommendation for each officer.
  • For most officers, position owners, and assignment teams, the number of options available in a single assignment cycle might not be large enough to necessitate personalized recommendations. However, ML-enabled recommendations are an enabling technology to facilitate a transition to a more flexible assignment system that selects officers for rotation according to development needs rather than assigning a set of officers who have been preselected for rotation.
  • ML-enabled recommendations augment, rather than automate, human decisionmaking. The design suggested in the report includes many points at which humans provide key inputs to inform ML recommendations. The virtue of a recommendation system is that it allows officers, position owners, and assignment teams to flexibly apply additional knowledge beyond the scope of the ML model.

Recommendations

  • The USAF should experiment with using a matching algorithm to combine officer and position owner preferences.
  • The USAF should experiment with delivering position and candidate recommendations using recommendation engines.
  • The USAF should gather new sources of data, such as officers' satisfaction with positions and position owners' satisfaction with officer performance.

Topics

Document Details

Citation

Chicago Manual of Style

Calkins, Avery, Monique Graham, Claude Messan Setodji, David Schulker, and Matthew Walsh, Machine Learning-Enabled Recommendations for the Air Force Officer Assignment System: Volume 5. Santa Monica, CA: RAND Corporation, 2024. https://www.rand.org/pubs/research_reports/RRA1745-5.html.
BibTeX RIS

Research conducted by

This research was prepared for the Department of the Air Force 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.