Naval Information Warfare Center - Atlantic, Charleston, SC

Laboratory Coordinator:
Josette Francis
P.O. Box 190022
North Charleston, SC 29419-9022
josette.francis@navy.mil

 

All participants must be U.S. citizens. NIWC ATLANTIC will not accept Permanent Residents or Dual Citizens as applicants for the ONR Summer Faculty Program.

Research Concentration Areas:

Artificial Intelligence, Machine Learning, and Data Science

  • Machine learning, deep learning, and transfer learning for computer vision, including for target detection, identification, and tracking
  • Natural language processing for question answering, semantic search, knowledge understanding, and in the presence of messy data
  • Reinforcement learning and optimization approaches for tactical decision making in delayed/disconnected, intermittent, and limited communications environments
  • Multi-criteria decision making, multi-objective optimization, and ensemble modeling methods
  • Explainable artificial intelligence, particularly for decision making, including explainable reinforcement learning
  • Physically realizable adversarial machine learning methods
  • Verification and validation of artificial intelligence and machine learning algorithms
  • Radiofrequency machine learning, fingerprinting, modulation classification, and pattern of life analysis
  • Edge processing for machine learning applications, including neuromorphic computing algorithms and hardware
  • Artificial intelligence and machine learning algorithms for quantum computing

 

Research Opportunities for SFRP 2022:

1: Title: Reinforcement Learning for Tactical Decision Making in Complex Environments
POC: 
Luke Overbey, lucas.overbey@navy.mil 

Task Description: This opportunity involves research and development of methods for developing optimal online policies for decision-making applications where agents are heterogeneous, distributed, and with varying levels of shared information. Research methods include utilization of deep reinforcement learning applications, as well as methods that increase explanability or allow for higher levels of abstraction of decision-making components/maneuvers. The desired end state would allow for agents to make adaptive decisions as new events occur or as the environment changes. The project will focus on one potential approach to solve such problems, partially informed by the area of expertise of the professor.

Place of performance: Naval Information Warfare Center Atlantic, Charleston, SC
Number of Professors Desired: 1

 

2: Title: Multi-Criteria Decision Making Methods for Course of Action Determination
POC: Luke Overbey, 
lucas.overbey@navy.mil 

Task Description: This opportunity involves research and development of methods for evaluating multiple decision-making models and determining an individual or set of courses of action based on these models that are Pareto optimal, given limited or imperfect information. In addition, we are interested in research involving methods to determine when machine learning and artificial intelligence models provide useful or beneficial solutions (contextual/environmental bounds for usefulness), as well as in what scenarios human intervention is warranted (human-in-the-loop vs human-on-the-loop). The project will focus on one potential approach to solve such problems, partially informed by the area of expertise of the professor.

Place of performance: Naval Information Warfare Center Atlantic, Charleston, SC
Number of Professors Desired: 1

 

3: Title: Realistic Adversarial Machine Learning Approaches
POC:
 Luke Overbey, lucas.overbey@navy.mil 

Task Description: This opportunity involves research and development of methods for developing adversarial machine learning approaches, applied to EO and IR image object detection and classification applications. We are particularly interested in methods that are based on or simulating realistic physical attacks within the scenes. These methods can involve synthetic data generation, three-dimensional modeling, or digital data manipulation approaches, and across varying levels of image qualities. The project will focus on one potential approach to solve such problems, partially informed by the area of expertise of the professor.

Place of performance: Naval Information Warfare Center Atlantic, Charleston, SC
Number of Professors Desired: 1

 

4: Title: Pattern of Life Analysis for RF Sensing Applications
POC:
 Luke Overbey, lucas.overbey@navy.mil 

Task Description: This opportunity involves research and development of methods for understanding patterns of life discoverable through individual or distributed radiofrequency sensors, based on information gleaned from signal type/modulation/identification information and across time and geographical components. Methods can include anomaly detection approaches and temporal or spatial pattern analysis using statistical or deep learning methods. The project will focus on one potential approach to solve such problems, partially informed by the area of expertise of the professor.

Place of performance: Naval Information Warfare Center Atlantic, Charleston, SC
Number of Professors Desired: 1

 

5: Title: Low-Shot and Transfer Learning Methods for Object Detection and Tracking in Full Motion Video
POC:
 Luke Overbey, lucas.overbey@navy.mil

Task Description: This opportunity involves research and development of methods for development of machine learning detectors and classifiers from overhead imagery with few examples. Methods can include low-shot approaches, siamese networks, transfer learning, or other related areas. In particular, we are interested in increasing accuracy for real-world applications where data is messy, out-of-focus, grainy, or in otherwise challenging environments. The project will focus on one potential approach to solve such problems, partially informed by the area of expertise of the professor.

Place of performance: Naval Information Warfare Center Atlantic, Charleston, SC
Number of Professors Desired:  1