Artificial Intelligence - Curriculum 388 (RES)
Program Officer
Brandon Holmes, LCDR, USN
Glasgow Hall East, Room E309
(831) 656-7980
brandon.holmes@nps.edu
Academic Associate
Duane Davis, Ph.D.
Glasgow Hall, Room 212
(831) 656-2733, DSN 756-2733
dtdavi1@nps.edu
Overview
This is a one-year resident Master of Science in Artificial Intelligence degree program, specifically designed for entrants with a strong foundation in computer science. It is exclusively focused on artificial intelligence (AI), in contrast to the Computer Science degree option with an artificial-intelligence specialization which requires 8 quarters. This intensive program aims to equip students with advanced AI techniques and skills tailored to address complex military challenges. Through a combination of challenging coursework, hands-on projects, and expert mentorship, students will gain a deep understanding of AI and its applications in areas such as defense systems, cybersecurity, surveillance, operations planning, and strategic decision-making. Graduates will emerge as proficient AI professionals capable of developing innovative solutions to enhance national security and defense capabilities.
Prerequisites
A baccalaureate degree in Computer Science or related field, with above average grades in mathematics, and at least one course in Artificial Intelligence. Applicants without an appropriate undergraduate degree that possess documented academic or practical experience in computer programming, discrete math, linear algebra, probability and statistics, computer operating systems and architectures, databases, and networking/distributed systems will also be considered, but may be required to take 12-week refresher quarter.
Degree Requirements
The degree of Master of Science in Artificial Intelligence is awarded after the satisfactory completion of a program which satisfies, as a minimum the following degree requirements.
1. At least 32 quarter hours of graduate level work, of which at least 16 quarter hours be at the 4000 level.
2. Completion of an acceptable thesis or capstone project.
3. To ensure a sufficient breadth across the field, the following courses represent the core that all students must complete:
•
CS4324, Adversarial and Secure Machine Learning (4-1)
•
CS4340, Trustworthy and Responsible AI (4-0)
Outcomes
- Present technical information about Al to technical and non-technical audiences, communicating complex data-related and Al-related concepts in a well-organized way through verbal, written, and/or visual means.
- Develop and recommend Al analytic approaches or solutions to problems and situations.
- Perform in and manage hands-on end-to-end team AI implementation projects with participation of subject-matter experts from the application domain.
- Describe the key kinds of Al professionals and their skills needed for Al projects.
- Analyze an AI task, and then apply data preprocessing and programming skills to put raw data into useful forms for Al.
- Describe the main types of Al models and how they work:
- Linear and nonlinear numeric models
- Probabilistic inference models
- Neural networks
- Logical models and inference
- Ontologies
- Heuristic search, planning, and time-series models
- Formal-game models
- Design, train, evaluate, and optimize machine-learning (ML) models by:
- Deep learning with convolutional neural networks and transformers
- Automated theorem proving
- Evolutionary methods
- Generative adversarial networks
- Ensemble learning methods
- Psychology-based methods
- Evaluate AI models:
- Calculate measures of accuracy for including recall, precision, and numeric fitting.
- Measure execution times and storage requirements.
- Evaluate explanations for AI models and their process of learning.
- Recognize problems of biased data and how to minimize them.
- Evaluate legal issues including privacy.
- Evaluate social and ethical implications of applications.
- Evaluate application of DoD ethics principles.
- Describe and apply modern concepts of software engineering to AI:
- Data operations and data warehouses
- DecSecOps (Development, Security, and Operations), early integration of software development, security protection, and software management
- MLOps (Machine Learning Operations), integration of machine and maintenance of its models
- Model-deployment tradeoffs
- Scalable AI, in and for CPU, GPU, HPC, and cloud-aware software systems
- Describe and analyze ways of thwarting threats from adversarial AI:
- Data manipulation by adversaries
- Security and cyber-security aspects of AI systems
- Assessing and mitigating risk in AI systems
- Analyze and apply AI to key military applications:
- Targeting
- Signal processing
- Intelligence gathering
- Autonomous systems
- Battle management
- Wargaming
- Cyberspace operations
- Predictive maintenance
- Logistics
- Automated help desks
- Generative language systems
Typical Course of Study
|
Refresher for students lacking background
|
CS3310 (4-1), Artificial Intelligence
|
CS3315 (3-1), Introduction to Machine Learning and Big Data
|
CS4000 (0-2), Harnessing AI
|
MA3333 (3-2), Math for AI
|
|
|
Quarter 1
|
CS4326 (4-1), AI on the Edge
|
CS4313 (3-2), Advanced Robotic Systems
|
CS4340 (4-0), Trustworthy and Responsible AI
|
CS4531 (3-2), Data Operations and DevSecOps
|
CS4904 (0-1), Current Research in AI
|
|
Quarter 2
|
CS4321 (3-2), Deep Learning
|
CS4333 (4-0), Current Directions in AI
|
CS4325 (4-1), Ontology and Theorem Proving for Trusted Systems
|
CS4327 (2-3), Naval AI Hackathon
|
|
|
Quarter 3
|
CS4323 (3-2), Bayesian Methods for Neural Networks
|
CS4317 (3-2), Language Systems
|
CS4330 (3-2), Computer Vision
|
Restricted Elective or CS0810 (0-8), Thesis
|
|
|
Quarter 4
|
CS4324 (4-1), Adversarial and Secure Machine Learning
|
MV4025 (3-2), Cognitive and Behavioral Models for Simulations
|
Restricted Elective or CS0810 (0-8), Thesis
|
CS0809 (0-8), Capstone Project in Computing or CS0810 (0-8), Thesis
|
|
Restricted Electives
The students could use the two Restricted Elective slots plus the Capstone slot for a thesis. Alternatively, students could use the two Restrictive Elective slots plus the Capstone slot for a shorter project. The Restricted Electives will be two related courses approved by the Curriculum’s Academic Associate. We require that the courses not significantly overlap existing courses in the Curriculum; CS4925 when titled “AI in War” would be an example of an acceptable Restricted Elective.
Refresher Quarter
| CS3310 | Artificial Intelligence | | 4 | 1 |
| CS3315 | Introduction to Machine Learning and Big Data | | 3 | 1 |
| CS4000 | Harnessing Artificial Intelligence | | 0 | 2 |
Quarter 1
| CS4326 | AI on the Edge: Enabling Physical Intelligence | | 3 | 1 |
| CS4313 | Advanced Robotic Systems | | 3 | 2 |
| CS4340 | Trustworthy and Responsible Artificial Intelligence | | 4 | 0 |
| CS4904 | Current Research in Artificial Intelligence | | 0 | 1 |
Quarter 2
| CS4321 | Deep Learning | | 3 | 2 |
| CS4333 | Current Directions in Artificial Intelligence | | 4 | 0 |
| CS4325 | Ontology and Theorem Proving for Trusted Systems | | 3 | 2 |
Quarter 3
| CS4323 | Bayesian Methods for Neural Networks | | 3 | 2 |
| CS4317 | Language Systems | | 3 | 2 |
| CS4330 | Introduction to Computer Vision | | 3 | 2 |
| CS0810 | Thesis Research | | 0 | 8 |
| -or- | | | |
| ELECT | Elective Course | | 0 | 4 |
Quarter 4
| CS4324 | Adversarial and Secure Machine Learning | | 4 | 1 |
| MV4025 | Cognitive and Behavioral Modeling for Simulations | | 3 | 2 |
| CS0810 | Thesis Research | | 0 | 8 |
| -or- | | | |
| ELECT | Elective | | | |
| CS0809 | Capstone Project in Computing | | 0 | V |
| -or- | | | |
| CS0810 | Thesis Research | | 0 | 8 |
Students could use the two Restricted Elective slots plus the Capstone slot as three thesis slots. Alternatively, students could use the two Restricted Elective slots for electives plus the Capstone slot for a shorter project. The Restricted Electives must be two related courses approved by the Curriculum's Academic Associate. We require that the elective courses not significiantly overlap existing courses in the Curriculum.