USMA Digital Commons (United States Military Academy, West Point)
Not a member yet
1355 research outputs found
Sort by
Counter-AI Tool System Design for AI System Adversarial Testing and Evaluation
No full textThis work consists of the initial recommendations and conclusions found while soliciting functional requirements for the research, design and development of a Counter-AI Tool for conducting adversarial testing and evaluation of artificial intelligence (AI) systems. The report includes a literature review of relevant AI concepts and extensive research within the adversarial AI domain. An intensive stakeholder analysis, consisting of requirement elicitation from over twenty governmental and non-governmental organizations, assisted in determining what functional requirements should be included in the system design of a Counter-AI Tool. The subsequent system architecture diagram takes user input, tests for various types of adversarial AI attacks, and outputs the vulnerabilities of the AI model. Prior to the operationalization of this tool, iterative experimentation will be conducted by partner organizations, which is the next step in the development and deployment of this Counter-AI Tool
A Monte Carlo Approach to Closing the Reality Gap
No full textWe propose a novel approach to the \u27reality gap\u27 problem, i.e., modifying a robot simulation so that its performance becomes more similar to observed real world phenomena. This problem arises whether the simulation is being used by human designers or in an automated policy development mechanism. We expect that the program/policy is developed using simulation, and subsequently deployed on a real system. We further assume that the program includes a monitor procedure with scalar output to determine when it is achieving its performance objectives. The proposed approach collects simulation and real world observations and builds conditional probability functions. These are used to generate paired roll-outs to identify points of divergence in behavior. These are used to generate {\it state-space kernels} that coerce the simulation into behaving more like observed reality.The method was evaluated using ROS/Gazebo for simulation and a heavily modified Traaxas platform in outdoor deployment. The results support not just that the kernel approach can force the simulation to behave more like reality, but that the modification is such that an improved control policy tested in the modified simulation also performs better in the real world
The Effectiveness of Virtual Reality for Training Human-Robot Teams
No full textHuman-Robot Teaming (HRT), or the collaboration of humans and robots to complete tasks, has become increasingly prevalent in the modern age. From medicine to military applications, robots have found a place in improving the capabilities of humans in challenging environments. As robots become incorporated in more aspects of human life, it is important that effective methods of training exist to prepare humans to interact with robot teammates. For this purpose, we created a tutorial for HRT using the Unity Game Engine. This tutorial is designed based on the analysis of previous literature and includes features found in other effective training systems and tutorials. Our tutorial is built on the framework of the Project Aquaticus HRT test-bed, a 3D, virtual reality (VR) simulation in which humans cooperate with autonomous robot teammates to compete in capture-the-flag in boats on the water. We introduce an experiment designed to compare the effectiveness of training via a 3D, VR tutorial to that of written manual instruction. This paper specifically introduces updates to our 3D simulation, tutorial, and proposed user studies to determine the system’s effectiveness at training humans for future human-robot teams
Sharing Cyber Capabilities within the Alliance - Interoperability Through Structured Pre-Authorization Cyber
Get PDFSharing cyber weapon/cyber capabilities requires trust between the member states, becoming a high-end policy decision due to the concerns of proliferation and the investment in designing a cyber-weapon that has a limited ’shelf-life’. The digital nature of cyber weapons creates a challenge. A cyber weapon can spread quickly, either self-propagating such as worms or via disclosure (and subsequent reuse) by malware researchers or malicious actors, raising proliferation concerns. Additionally, a cyber-weapon can be copied by the adversary or reverse engineered. Once the weapon is released, the adversary will eventually address the vulnerability, and the opportunity is gone. These factors raise the threshold between member states to share cyber weapons and cyber capabilities. Alliances, like NATO, prepare for a unified multinational, multi-domain fight; meanwhile, the national cyber forces are still operating as solitaires with limited interoperability and sharing. There is a need in the collective defence posture to integrate the multinational cyber force to achieve interoperability
Identification of Latent Structure in Spatio-Temporal Models of Violence
No full textThe modeling of violence, including terrorist activity, over space and time is often done using one of two broad classes of statistical models. Typically, the location of an event is modeled as a spatio-temporal point process and the latent structure is either modeled through a latent Gaussian process motivated by a log-Gaussian Cox process or through data dependency similar to a Hawkes process. The former is characterized through dependence in an unobserved latent Gaussian process, while the later assumes a data driven dependence in the data. While both techniques have been used successfully it remains unclear whether the processes are practically different from one another. In this manuscript, we demonstrate that in many situations, the most common statistic to characterize clustering in a process, Ripley\u27s K function, cannot differentiate between the two processes and should not be used
A Troubling Silence on Prisoners of War
No full textThe Russian-Ukrainian War is the most significant ground war in Europe since World War II and a test of the treatment of prisoners of war (PoWs). With a few exceptions in the Balkans in the 1990s, Europe has not experienced a large number of captured personnel since 1945. The American experience is close to50 years away, ending with the War in Vietnam. Years of counterinsurgency operations have bred complacency; the notion that potentially thousands of soldiers could become prisoners of war (PoWs) has, until the events in Ukraine been seen as unlikely. Preparing soldiers for captivity, confinement, and interrogation has also lapsed, which risks mistakes and unnecessary harm in a conflict
Enhancing Military Training Using Extended Reality: A Study of Military Tactics Comprehension
No full textThis study identifies that increasing the fidelity of terrain representation does not necessarily increase overall understanding of the terrain in a simulated mission planning environment using the Battlefield Visualization and Interaction software (BVI; formerly known as ARES (M. W. Boyce et al., International Conference on Augmented Cognition, 2017, 411–422). Prior research by M. Boyce et al. (Military Psychology, 2019, 31(1), 45–59) compared human performance on a flat surface (tablet) versus topographically-shaped surface (BVI on a sand table integrated with top-down projection). Their results demonstrated that the topographically-shaped surface increased the perceived usability of the interface and reduced cognitive load relative to the flat interface, but did not affect overall task performance (i.e., accuracy and response time). The present study extends this work by adding BVI onto a Microsoft HoloLens™. A sample of 72 United States Military Academy cadets used BVI on three different technologies: a tablet, a sand table (a projection-based display onto a military sand table), and on the HoloLens™ in a within-subjects design. Participants answered questions regarding military tactics in the context of conducting an attack in complex terrain. While prior research (Dixon et al., Display Technologies and Applications for Defense, Security, and Avionics III, 2009, 7327) suggested that the full 3D visualization used by the Hololens™ should improve performance relative to the sand table and tablet, our results demonstrated that the HoloLens™ performed relatively worse than the other modalities in accuracy, response time, cognitive load, and usability. Implications and limitations of this work will be discussed
Operations Research (In Mathematics in Cyber Research)
No full textIn the last eighty years, the discipline of operations research has been used extensively to provide analytical evidence for research outcomes in the mathematical sciences. Problems of interest may involve variations of cost-benefit analyses, the measurement of attributes or features, differing levels of uncertainty, and constrained or unconstrained frameworks. The analyses are not always concerned with generating unique optimal solutions; often, an array of near-optimal solutions may be more appropriate depending on the assumptions, factors, and objectives of a particular study. Research work may also be more aligned toward investigating how complex processes behave either in a static or dynamic manner. Sensitivity analyses may accompany outcomes to provide an indication of the robustness of different solutions. In summary, the assortment of methods, tools, and analytics in operations research is vast and growing. Emerging fields such as data science and techniques such as machine learning will fuel continued growth in this mathematical discipline for many years to come.
Interestingly, the domain of cyber theory and applications is also growing, and the products and processes in this realm have qualities and characteristics that warrant using methods in operations research. Cyber attacks can be modeled using hierarchical threat structures and may involve competing decision strategies from both an organization or individual and the adversary. Network traffic flow, intrusion detection and prevention systems, interconnected human-machine interfaces, and automated systems all require higher levels of complexity in modeling and possess inherently random sub-processes. Attributes, such as cyber resiliency, network adaptability, security capability, and information technology flexibility require the measurement of multiple characteristics, many of which may involve both quantitative and qualitative interpretations. For nearly every organization that is invested in some cybersecurity practice, decisions must be made that involve the competing objectives of cost, risk, and performance.
This chapter will highlight only a small portion of the operations research space. The sub-discipline of decision analysis is first described, which may involve the modeling or measuring of actions, factors, or responses using mathematics to aid in decision making. The sub-discipline of mathematical optimization is then discussed, which may involve solving for the most desirable outcome or solutions given a constrained or unconstrained framework. The last broad topic, techniques in stochastic process modeling are provided to account for the random or uncertain behavior of cyber processes and where insight and greater forecasting power may be achieved. Along the way, the focus will be on the mathematics associated with these sub-disciplines of operations research. Finally, at the conclusion of this chapter, research implications and extensions are offered to the examiner that desires to carry these practices forward in cyber research.https://digitalcommons.usmalibrary.org/aci_books/1017/thumbnail.jp
Accounting for Remote Testing in West Point’s Advanced Math Program
Get PDFWhen new students first arrive at the United States Military Academy (USMA), they soon find out what classes they will take during the upcoming Fall semester. Some cadets with strong mathematical backgrounds discover they have been placed into the Advanced Mathematics Program (AMP). Over the years, AMP course directors have used a variety of techniques to identify the best candidates for AMP which usually figures out to the top 30% of the incoming class. Most of these techniques considered similar criteria ranging from USMA’s own college entrance examination rank (CEER) score, which factors in SAT/ACT scores among several other scores, to their Fundamental Concepts Exam (FCE) and Calculus Placement Exam (CPE) scores. Recently, the FCE and CPE have both been remotely administered during the Summer before students attend USMA to allow for expedited placement. In this paper, we use placement and post-placement scores to measure the effectiveness of AMP’s current tier-based placement system. We found that the system results in few errors and can be easily optimized further through additional quality control measures
A Celebration of West Point Authors
Get PDFToday we celebrate the 480 works of scholarship produced at the Academy between January and June 2022. The program features speakers whose work centered around civil-military relations.https://digitalcommons.usmalibrary.org/books/1064/thumbnail.jp