Phantom Fellowship Program Culminates in BAH Optimization Project

  • Published
  • By DAF AI Accelerator Public Affairs
  • Department of the Air Force AI Accelerator

Last month, the members of Cohort 6 completed their Phantom Fellowship Program with the Department of the Air Force-Massachusetts Institute of Technology Artificial Intelligence Accelerator which culminated in a capstone project that examined possible solutions using AI and Machine Learning models to support and augment the process of calculating Basic Allowance for Housing. 

The Phantom Fellowship is a rigorous program for Airmen and Guardians interested in gaining exposure to AI/ML technologies. Selected “Phantoms” are temporarily embedded with the DAF-MIT AI Accelerator, an organization dedicated to developing and deploying difference-making AI/ML capabilities by guiding fundamental, dual-use research in pursuit of the DAF’s most challenging problems. During the fellowship, Phantoms are charged to augment the AIA’s research efforts, continue their self-led AI/ML education, research and write a paper on the impacts of AI/ML to their home units, and work together on a capstone project. Cohort 6 consisted of civilians, officers and enlisted members from active duty and the national guard, and even included participants from the Department of the Navy. Phantoms come from an array of career fields to include loadmasters, analysts, personnelists, program managers and software engineers. Phantom cohorts are chosen from several hundred applicants and fill both technical and non-technical roles to accelerate the DoD-wide adoption and understanding of AI. 

The capstone project was a Phantom led and organized portion of the fellowship program. Serendipitously, while Cohort 6 Phantoms were brainstorming on their project topic in late September, the Department of Defense announced the temporary increase of BAH rates for 28 military housing areas across the United States. The authorization resulted from unusually large spikes in the median rental housing costs and leaders suggested looking to AI for help with reviewing military compensation during volatile financial times. Senior Airman Alexander Carey, a Phantom and Scientific Applications Specialist from the Air Force Technical Application Center at Patrick SFB, proposed the cohort delve into the potential application of AI to the BAH process for their capstone project. “Typically when we hear about AI/ML, it is being used as a buzzword in the media or as a selling point for some new product, process, or service,” said Carey. “Done correctly, AI/ML can accelerate or enhance jobs ranging from data analytics to supply chain management and even potentially recruiting. My aim was to determine if a combination of publicly available data and AI/ML processes could be used to enhance the current BAH calculation process, hopefully to better reflect the state of the housing market for the benefit of all service members.” Alexander was later selected to receive the AIA’s Director Award for his outstanding contributions to the program and significant personal growth during the fellowship. 

Advancing forward with a compelling topic, the Phantoms utilized their backgrounds and strengths to devise an execution plan for the capstone project during the final 2 months of their fellowship. Capt. Zach Baumann, a Phantom and Force Support Officer from Air Force Personnel Center at Randolph Air Force Base, TX, stepped forward to lead the team.

During a previous assignment at Headquarters, U.S. Forces Korea, Baumann worked with the Office of Secretary of Defense Military Compensation Policy to evaluate cost of living allowances for all service members residing in the Republic of Korea. Calling on his prior connections, Baumann pitched the proposed capstone topic and garnered a working partnership with their office to receive direction on potential areas that could utilize AI in their processes. With this crucial guidance and a solid problem set, Cohort 6 got to work.

“I was excited OSD was willing to hear what we could come up with,” said Baumann. “I know BAH will always be a hot topic, and it seems that there’s no shortage of opinions out there about it. We learned quickly that calculating BAH for over 1.1 million members across 300 military housing areas with annual BAH payments exceeding $25B is no small task. So after humbling ourselves, we entered the project with a very practical mindset; can we utilize our fresh perspectives and some modern, data-driven tools to make some parts of BAH a little bit better?”

After extensively researching the current practices for calculating BAH, the Phantoms organically divided into 3 cross-functional teams to each examine different challenges faced by the OSD team. One team researched and developed a prototype model for an optimization tool to monitor real-time market supply and demand, an angle that is not currently accounted for in the housing allowance process. Another team sought to replicate BAH calculations using only public data from the U.S. Census and Consumer Price Index to either replace or augment the current 7-month, time-intensive data collection processes involved in BAH rate setting. The last team web scraped similar housing market data in an effort to provide BAH validation checkpoints at any time the markets unexpectedly spike.

“It was an awesome experience to collaborate with such a diverse group of Phantoms from all over, all bringing their strengths,” said Baumann. “From the technical to the project managers to all of us just giving and receiving feedback, this has been one of those times where the whole was greater than the sum of the parts. But of course, we had some wicked smart people.”

On December 8, Cohort 6 presented their solutions to OSD’s Military Compensation Policy team. This team included Mr. Kevin Scott, an OSD economist and housing program advisor, who served as the Phantom liaison for the capstone project. “This project has really helped to distill and focus the conversation around potential advancements in BAH program design. The Phantom Team developed innovative concepts the Department can build upon to more fully leverage data-driven automation in BAH processes. I was impressed with the team’s commitment and eagerness to help tackle some of the biggest challenges in modern housing market estimation.”

Part of the Phantom Fellowship Program mission is to develop advocates of AI within the DoD to help the rest of the services understand the capabilities and limitations of AI. “AI/ML technologies are continually advancing, and it has been enlightening to be involved in that process during my time in the fellowship,” said Carey. “Our capstone project demonstrated that AI/ML solutions are not necessarily beyond the ability of the average person. The models and algorithms we used in our project were nothing new, mostly stemming from our prior experiences or self-driven learning during the fellowship. Classes are available to everyone in the DOD, and I encourage anyone interested to take them. I believe that determination and passion are a larger part of the battle to develop new skills than raw talent, and that everyone has an important role in bringing AI to the DoD.” 

“I am very proud of our Phantom program, and in this Cohort in particular, who took on a very challenging problem in their capstone,” said Col. Garry Floyd, DAF-MIT AIA Director. “There are AI use-cases everywhere across DoD and the DAF. If we can become better masters of our data through AI, we can not only enhance lethality in combat operations, but also improve the overall experience of those who serve.” 

To learn more about the Phantom Fellowship Program and other education opportunities like Digital University, visit here