STRATEGIC PARTNERS

AHFA partners with multiple organizations who share similar missions to empower and engage youth in aviation. A complete list of our partners include:

AHFA

ELIGIBILITY REQUIREMENTS

  • Be at least 16 years old by 1 June of the flight training year.
  • Be enrolled in high school* or homeschool during the flight training year.
  • Be a US citizen **
  • Have no more than 5 powered flight hours 
  • 3.0 current Grade Point Average either with an unofficial GPA Transcript or a registrar/counselor verifiable memorandum.

Notes: 

  • No flight or aviation experience necessary (we teach you to fly!)  
  • No commitment or obligation to the Air Force  
  • *USAFA & AFROTC Cadets may apply through their institutions
  • **Non-US Citizens regardless of residency status are ineligible to apply

 

Videos from DVIDShub.net

Michael Robinson - Topological Features in Large Language Models (and beyond?)
Air Force Research Laboratory
Video by Kenneth M McNulty, Kevin D Schmidt
Oct. 25, 2024 | 01:00:53
In this edition of QuEST, Michael Robinson will discuss topological features in large language models

Key Moments and Questions in the video include:
Acknowledgement of colleagues from DARPA and Galois
Manifolds in machine learning
LLM token space is higher dimensional
Manifold spaces tend to be negatively curved
LLM turn text into vectors
Transformers turn vectors into new text
How do we turn the text into vectors?
We think of LLM as being trained on all human language, but they have not
GPT2 Open source LLM as the source for model
ChatGPT2 used as the example
Tokens have topology and geometry
Words are a categorical variable
Vectors are a numerical variable
Mixing data types can lead to some problems
Why care about the token space?
Not all tokens correspond to a valid vector
Estimating dimensions
Volume of a sphere
Log of Volume vs log of radius curves
Ricci scalar curvature
Stratifications are visible
GPT2 uses a state space that is not a manifold
Dollar sign shown different in GPT2 because the $ is used in code where other currency symbols are not
GPT2’s 768 dimensions unwrapped using tSNE
Tokens with leading spaces
Beginnings of words show up in separate piece of low dimension
Visual similarity to hyperbolic plane
LLEMMA7B dimensions
Plotting dimension
Dark space are non-printing characters
Thinking about how neural activation patterns work
We have been thinking about manifold learning out of mathematical convenience
State spaces are not manifolds
Open presentation to conversation More

Tags

questAFRLACT3large language modelsmanifolds in machine learningArtificial Horizon Indicator (AHI) More
AETC Commander's Challenge
Play Button
1:22
AETC Commander's Challenge
Feb 2025 LPQ&A featuring Mr. Jim Goodell presenting LPQ&A Learning Engineering Level-Setting
Play Button
53:25
Feb 2025 LPQ&A featuring Mr. Jim Goodell presenting LPQ&A Learning Engineering Level-Setting
LPQ&A 28 January 2025 featuring Dr. Stephanie Moore presenting Authentic Assessments: Designing Assessments for Learning in any Modality
Play Button
52:09
LPQ&A 28 January 2025 featuring Dr. Stephanie Moore presenting Authentic Assessments: Designing Assessments for Learning in any Modality
Global College of Professional Military Education: CMSAF Flosi
Play Button
2:10
Global College of Professional Military Education: CMSAF Flosi
Now Playing
Michael Robinson - Topological Features in Large Language Models (and beyond?)
Michael Robinson - Topological Features in Large Language Models (and beyond?)
LPQ&A Sept 2024
Play Button
59:38
LPQ&A Sept 2024
LPQ&A August 2024
Play Button
48:37
LPQ&A August 2024
LPQ&A Jul 2024 featuring Mr. Nick Floro presenting Upskilling Your Designs: Skills, Tools, and Techniques
Play Button
01:00:29
LPQ&A Jul 2024 featuring Mr. Nick Floro presenting Upskilling Your Designs: Skills, Tools, and Techniques
LPQ&A 27 Jun 2024 featuring Mr. Josh Cavalier presenting ChatGPT prompt strategies for Learning and Development
Play Button
01:03:30
LPQ&A 27 Jun 2024 featuring Mr. Josh Cavalier presenting ChatGPT prompt strategies for Learning and Development
TEASER – Airman Magazine: Artificial Intelligence
Play Button
2:13
TEASER – Airman Magazine: Artificial Intelligence
3rd annual DAF Learning Symposium
Play Button
2:09
3rd annual DAF Learning Symposium
USAF Test Pilot School and DARPA announce breakthrough in aerospace machine learning
Play Button
0:56
USAF Test Pilot School and DARPA announce breakthrough in aerospace machine learning
Responsible AI Development
Play Button
4:59
Responsible AI Development
MSOTL Promotional Video
Play Button
1:28
MSOTL Promotional Video
Dr. Raj Sharma and Jasper Craig
Play Button
55:05
Dr. Raj Sharma and Jasper Craig
See More

Resources

AHFA Locations & Training Partners

  • California Aeronautical University, CA 
  • California Baptist University, CA 
  • Marion Military Institute, AL 
  • Oklahoma State University, OK  
  • South Dakota State University, SD 
  • Troy University, AL
  • Schreiner University, TX
  • University of Texas San Antonio, TX
  • Tennesse State University, TN

*To learn how to become one of our training locations please email: 

 Afrs.ahfa.studentapplications@us.af.mil 

SCHOLARSHIP RECIPIENTS RECEIVE* 

  • Up to 12-15 flight hours 
  • Housing and meals during training 
  • Transportation to/from training location 
  • Classroom training (ground school) 
  • Flight simulator training 
  • All training is provided by FAA Certified Flight Instructors 
  • Access to university recreation facilities 
  • Mentorship from Air Force aviators  

*All items funded by USAF except:

  • FAA Class III Med Certificate
  • Luggage during travel
  • Personal driving to/from university assigned session

Contact us

 

 

                                                              Please direct program questions to: Afrs.ahfa.studentapplications@us.af.mil