AI and Autonomy Workshop

2020-27-29-October-AI-and-Autonomy-Workshop-Laurel-MD

11100 Johns Hopkins Rd.
Laurel, MD 20723

Register
Add to Calendar

Overview

Terms of Reference (TOR) document as of 02 March.

Within the next five years, the DoD will start to deliver AI-enabled capabilities and autonomous systems (e.g., the Army’s Expedient Leader-Follower or ExLF). Much like when the “Big 5” (M1, M2, Multiple Launched Rocket System, Apache, and Blackhawk) that were fielded in the 1980’s, for AI and autonomous systems, analysts will be asked to conduct cost analysis, force structure analysis, and analyze changes in Tactics, Techniques, and Procedures. Thus, analysts must be familiar with capabilities and limitations of AI and autonomy technologies. Like the 2018 AI and Autonomy I Special Meeting, the October 2020 meeting is another step in the journey to build this knowledge set and give analysts a look into the crystal ball of possible AI and autonomous capabilities that will probably mature within the next five years. Sessions include:

  • Defense and Security Applications – Andrew Ng has estimated that any task requiring less than one second of mental thought can be implemented by AI/autonomous systems.[1] The Defense and Security Applications session will provide evidence as to whether this is accurate in DoD and Security mission spaces. The session will highlight specific use-cases, unique perspectives, rules of thumb, experienced difficulties and insights regarding what it takes to develop and deploy these capabilities.
     
  • Ethics and Explainability – AI is deepening our world in new ways and creating opportunities for improvements in Defense and other sectors. AI’s apparent blackbox nature is however, fueling debates around ethics and explainability in increasingly complex ways. Issues such as bias, fairness, privacy, trust, and transparency are ubiquitous, and necessitate further conversations around the role of ethics and explainability in the design, implementation, operationalization, and regulation of AI. This session will present current national and international concerns in this arena to motivate efforts towards more explainable and ethical AI frameworks.
     
  • Test and Evaluation Capabilities – This session will provide a primer on the challenges and existing research being done in the domain of test and evaluation of autonomous and AI-based systems. It will address the various ways in which low level performance evaluation and benchmarking of these systems are supported, and how these low level metrics and techniques can be connected to higher level operational goals to determine the overall effectiveness and operational utility of autonomous and AI-based systems.
     
  • Ongoing International and Industrial Applications – This session addresses applications and research in other countries and domains that may advance the technology base for military AI and autonomy. Of specific interest are ‘bridging’ domains supported by Operations Research (OR), such as mission planning and multidomain integration.[2]
     
  • Looking Five Years into the Future – AI developments in the next five years, if implemented correctly, will continue to improve military operations. What’s unknown is how decisive an advantage that will be and whether the gap between our near peers is widening or closing. This session will explore where we are across a number of dimensions, including autonomy, neural networks, and computer vision.
     

 Participants from government, industry, academia, local and international are welcome to attend. To foster participation from a broad spectrum of analysts and researchers interested in AI and autonomy, this Workshop will be held at the unclassified level.

If you are interested in assisting or making a presentation, please contact Dr. Nate Bastian (nathaniel.bastian@fulbrightmail.org) and Mr. Don Timian (dontimian@hotmail.com).

[1] https://www.wsj.com/articles/how-artificial-intelligence-will-change-everything-1488856320

[2] Continuing this theme from MORS’ first AI and Autonomy Special Meeting, but in the narrower context of the military sponsored research.

Agenda

Tutorials:

  • “Approachable AI”; Instructor – Dr. Nate Bastian; This tutorial surveys several topics in artificial intelligence engineering with the overall objective of making the AI discipline more approachable for not only leaders and practitioners but also analysts, engineers, and scientists. In this tutorial, we will 1) highlight the latest trends and applications in AI, 2) review the traditional and emerging AI technologies through the lens of data, tools and solution techniques, 3) describe AI product development, including both the engineering and management of AI-enabled solutions, and 4) discuss the impacts of AI on the DoD and society at-large.
     
  • “Creating AI Enabled Systems”; Instructor – Dr. John Piorkowski; Achieving the full vision of AI requires a system perspective to effectively leverage algorithms, data, and computing power. Creating AI-enabled systems includes thoughtful consideration of the mission decomposition for AI solutions, engineering data for algorithm development, and deployment strategies. To realize the impact of AI technologies requires a systems perspective that goes beyond the algorithms. The objective of this course is to bring a system perspective to creating AI-enabled systems.
     
  • “eXplainable AI”; Instructor – Dr. Stanton Price; Many state-of-the-art machine learning methodologies have a black-box nature – you do not know what is going on inside; you give the black-box (trained deep learning model) input data, it processes the data, and gives back an output without any explanation behind the decision/prediction that was made. Developing highly effective deep learning models that are also interpretable and/or provide some level of intuition into how its decisions are made is becoming an essential necessity for most DoD and Department of Homeland Security (DHS) applications seeking to employ AI-based technologies into theater. This tutorial will focus on the latest research in the area of explainable AI, with a focus on machine learning-based computer vision applications relevant to the DoD (Intelligence, Surveillance, and Reconnaissance (ISR), targeting, etc.). Further, demos will be presented using MATLAB to illustrate practical use cases and highlight their strengths, weaknesses, and areas of open research.
     
  • Data Cleaning and Knowledge Management; Instructor – Dr. Jim Kyung-Soo Liew; This tutorial will review several real-world challenges and solutions with data cleaning, gathering, managing, and storing knowledge.  Participants will be actively engaged in a lively discussion reviewing case-studies filled with practical problems, obstacles, and solutions.  If data is the fuel for any successful AIproject, getting the data correct is imperative, but how correct does it have to be to drive Return on Investment (ROI)?  Topics covered include:  1) Cleaning Data-The Basics, 2) Extracting Value from Data, 3) Practical Strategies in Storing and Sharing Knowledge, and 4) Scaling Data Solutions to Fuel AI Success.

If you are not familiar with Machine Learning, recommend you attend Dr. Nate Bastian’s “Approachable AI” tutorial, during the morning of 26 October, then attend – in the afternoon – Dr. Stanton Price’s “eXplainable AI” tutorial.

Registration

Payment Options

The "Invoice Me" option can only be used for the registration fee. If you need assistance registering please contact Ms. Tina Yan, tina.yan@mors.org

Non-Member
  
AI 3 Day Meeting
  
Member Non-Member
US Federal Government MORS Sponsor
Includes Optional Tutorial
$450 $550
US Federal Government MORS Sponsor
Tutorial Only
$100 $150
U.S. Federal Government
Includes Optional Tutorial
$500 $600
U.S. Federal Government
Tutorial Only
$125 $175
National Research Partner(IDA Only)
Includes Optional Tutorial
$523 $618
National Research Partner(IDA Only)
Tutorial Only
$125 $175
All Others
Includes Optional Tutorial
$550 $650
All Others
Tutorial Only
$150 $200

JHU/APL participants may contact Ms. Tina Yan, tina.yan@mors.org for discounted registration fees for JHU/APL staff only.

Security

Who can attend (Clearance Requirements):

This meeting is open to U.S. citizens and FVEY Partners (Australia, Canada, New Zealand, UK). Non U.S. citizens of U.S. allies may attend at the discretion of MORS. Non U.S. attendees will be notified if their registration to attend is not accepted.

Lodging

Homewood Suites Columbia/Laurel
Rate: $108
Address: 7531 Montpelier Road
Laurel, MD 20723
Phone: (240) 360-2725