IoT for Defense and National Security

8Principles of Robust Learning and Inference for IoBTs

Nathaniel D. Bastian¹, Susmit Jha², Paulo Tabuada³, Venugopal Veeravalli⁴, and Gunjan Verma⁵

¹Army Cyber Institute, United States Military Academy, West Point, NY, USA

²Neuro‐symbolic Computing and Intelligence, CSL, SRI International, Menlo Park, CA, USA

³ECE Department, University of California at Los Angeles, Los Angeles, CA, USA

⁴ECE Department, University of Illinois at Urbana‐Champaign, Champaign, IL, USA

⁵U.S. Army DEVCOM Army Research Laboratory, U.S. Army Futures Command, Austin, TX, USA

Abstract

The Internet of Battlefield Things (IoBTs) operate in an adversarial rapidly‐evolving environment, necessitating fast, robust and resilient decision‐making. The success of machine learning, in particular deep learning methods, can improve the performance and effectiveness of IoBTs, but these models are known to be brittle, untrustworthy, and vulnerable. In this chapter, we discuss the principles and methodologies to make machine learning models robust, resilient to adversarial attacks, and more interpretable for human‐on‐the‐loop decision‐making. We also identify the key challenges in developing trustworthy machine learning for IoBTs.

8.1 Internet of Battlefield Things and Intelligence

The Internet of Battlefield Things (IoBTs) [1, 2] aims at providing a pervasive, heterogeneous sensing and actuation capability to enhance command and control system autonomy and agility, information analytic capabilities against adversarial ...