USMA Digital Commons (United States Military Academy, West Point)
Not a member yet
1355 research outputs found
Sort by
Cross-platform Information Spread During the January 6th Capitol Riots
No full textSocial media has become an integral component of the modern information system. An average person typically has multiple accounts across different platforms. At the same time, the rise of social media facilitates the spread of online mis/disinformation narratives within and across these platforms. In this study, we characterize the coordinated information dissemination of information laden with mis- and disinformation narratives within and across two platforms, Parler and Twitter, during the online discourse surrounding the January 6th 2021 Capitol Riots event. Through the use of username similarity, we discover joint theme endorsements between both platforms. Using anomalously high volume of shared-link matches of external websites and YouTube videos, we discover separate information consumption habits between both platforms, with very few common sources of information between users of the different platforms. However, through analyzing the similarity of the texts with Locality Sensitive Hashing of constructed text vectors, we identify similar narratives between the platforms despite separate consumption of external websites, highlighting the similarities and differences of information spread within and between the two social media environments
Using a Case Study to Teach Leaders How to Enact Positive Organizational Change
Get PDFhttps://digitalcommons.usmalibrary.org/books/1062/thumbnail.jp
Predicting Phishing Vulnerabilities Using Machine Learning
No full textThis paper examines the ability to use machine learning to predict an undergraduate student’s actions upon receiving a phishing email. The machine learning models used in this work were trained with actual phishing results augmented with student’s background and administrative data. The ultimate goal of this project is to better identify members of an organization that are at risk from phishing, to provide targeted cyber security training. This targeted training will increase the security posture of an organization and minimize unnecessary training and productivity loss. The results of multiple machine learning techniques demonstrate that this approach is viable with validation accuracy ranging from 49 to 86%. Other metrics are used to evaluate the viability of the approaches, recall is determined to be the most important. The model with the best performance in validation using these two metrics was a Support Vector Machine (SVM). The SVM approach was able to predict whether a cadet would be compromised upon receipt of a phishing attack with a 55% accuracy while maintaining a recall score of 71%. When using the trained model on new data after training and validation the Logistic Regression model had the highest performance, accurately predicting whether a cadet would be compromised upon receipt of a phishing attack with a 86% accuracy while maintaining a recall score of 16%
Ethical Frameworks for Cybersecurity: Applications for Human and Artificial Agents
No full textAnalogical reasoning is one of the most powerful tools that humans have to understand their environments. Novel situations can be rendered familiar by parallels with culturally available schemata and more fundamental exchange norms such as reciprocity. The use of analogies such as “cyberwar” create trade-offs in that they both emphasize and deemphasize the affordances of a situation or a technology. Although they might reflect the approaches of analysts and policymakers, they might not accurately reflect the understanding of users. Instead of specific analogies, we instead suggest that general schemata (e.g., social dilemmas) provides a more agnostic basis for understanding user behavior. Moreover, in that human social agents use these schemata, they can be incorporated into autonomous and intelligent systems to simulate human behavior and can also be used to detect irregular network activity that violates the exchange norms humans typically use.https://digitalcommons.usmalibrary.org/aci_books/1019/thumbnail.jp
Deep Blue Wants You: Identifying and Addressing Sources of Bias in AI Systems to Support Human Resources Decisions
No full textDue to benefits like increased speed, Artificial Intelligence (AI)/Machine Learning systems are increasingly involved in screening decisions which determine whether or not individuals will be granted important opportunities such as college admission or loan/mortgage approval. To discuss concerns about potential bias in such systems, this chapter focuses on AI to support human resource decisions (e.g., selecting among job applicants). As AI systems do not inherently harbor prejudices, they could increase fairness and reduce bias. We discuss, however, that bias can be introduced via: (i) human influence in the system design, (ii) the training data supplied to the system, and (iii) human involvement in processing system recommendations. For each of the above factors we review and suggest possible solutions to reduce/remove bias. Notably, developing AI systems for screening decisions increased scrutiny for bias and raised awareness about pre-existing bias in human decision patterns which AI systems were trained to emulate.https://digitalcommons.usmalibrary.org/aci_books/1018/thumbnail.jp
Reference Percentiles for Bioelectrical Phase Angle in Athletes.
No full textThe present study aimed to develop reference values for bioelectrical phase angle in male and female athletes from different sports. Overall, 2224 subjects participated in this study [1658 males (age 26.2 ± 8.9 y) and 566 females (age 26.9 ± 6.6 y)]. Participants were categorized by their sport discipline and sorted into three different sport modalities: endurance, velocity/power, and team sports. Phase angle was directly measured using a foot-to-hand bioimpedance technology at a 50 kHz frequency during the in-season period. Reference percentiles (5th, 15th, 50th, 85th, and 95th) were calculated and stratified by sex, sport discipline and modality using an empirical Bayesian analysis. This method allows for the sharing of information between different groups, creating reference percentiles, even for sports disciplines with few observations. Phase angle differed (men
Army Officer Corps Science, Technology, Engineering and Mathematics (STEM) Foundation Gaps Place Countering Weapons of Mass Destruction (CWMD) Operations at Risk – Part 2
No full textThis is the second of three articles from the authors describing the risk to Joint Operations incurred by an Army that is vulnerable to the STEM challenges faced in a great power competition involving CWMD operations. In Part 1, we described the problem: “The Army’s failure to emphasize STEM competence in the Army officer corps outside of Functional Areas creates risk to mission accomplishment in CWMD multi-domain operations. The Army must prioritize STEM education in accessions and throughout PME to prepare commanders for effective science and technology (S&T) informed decision making within mission command in CWMD multi-domain operations”. For Parts 2 and 3, we utilize the Joint Operational Model, Notional Phasing for Predominant Military Activities, from JP 3-0, Joint Operations, to describe the risk of an Army officer corps lacking STEM dominance for CWMD operations during a regional or great power competition involving CWMD operations. In this article, we address the risk of our current efforts as we operate in Phase 0 (Shape) and Phase 1 (Deter) while our final article (Part 3) will examine the transition to decisive action / unified action with Phase 2 (Seize the Initiative) through Phase 5 (Enable Civil Authority)
Holmquist-Johnson-Cook Constitutive Model Validation and Experimental Study on the Impact Response of Cellular Concrete
No full textIn a previous study by Davis and Dequenne, a Holmquist-Johnson-Cook (HJC) constitutive model for a cellular concrete with a nominal density of 1442 kg/m3 was developed from existing direct tension, uniaxial strain, and triaxial shear testing conducted at the United States Army Corps of Engineers Engineer Research and Development Center (ERDC) and Sandia National Laboratory (SNL). The resulting constitutive model was compared to depth of penetration results from testing conducted by Goodman at the Aberdeen Test Center with promising results. This study seeks to build on this previous work by producing depth of penetration and perforation experiments using non-deforming projectiles into a similar cellular concrete for validation of the fit HJC model. Depth of penetration experiments were conducted by firing into a 305 mm thick panel over a velocity range of 200–800 m/s with the strike velocity and depth of penetration recorded for each experiment. Perforation experiments were conducted over a range of 200–800 m/s against panels with thicknesses of 38 mm, 76 mm, and 114 mm with the strike velocity, residual velocity, and crater characteristics recorded for each experiment. 2D numerical simulations were conducted for each experiment and the results were compared for initial model validation, but additional experimental testing and simulation is required. There is error between the experimental and numerical results and a sensitivity analysis should be conducted to determine where additional testing is appropriate to improve the model’s correlation with experimental results
Constitutive Modeling and Validation of Sintered Metal Powders Subjected to Large Strains and High Strain Rates
No full textThe development of advanced small caliber weapon systems has resulted in rounds with more material penetration capabilities. The increased capabilities may mean that existing live-fire facilities will no longer be adequate for the training and certification of military and law enforcement personnel, which could result in training constraints and possibly expensive upgrades to improve the safety of existing facilities. New training ammunition manufactured from novel structural materials are needed to allow for the safe, continued use of live-fire shoot house facilities. The goal of this project is to characterize a sintered metal powder and fit a suitable constitutive model for simulation in support of numerical design. A pressed and sintered blend of copper-tin was selected as a suitable representative material for this application. Samples were tested in uniaxial compression under quasi-static conditions and elevated temperatures. Dynamic compression testing at strain rates up to approximately 105 s−1 was conducted using a split-Hopkinson bar. The results of these tests were then used to fit Johnson-Cook and Zerilli-Armstrong strength models to the test data. The models were fit by selecting points from test data at different strain rates and elevated temperatures. This system of equations was then solved for each model while using the same test data to ensure a fair comparison of the results. A Mie-Gruneisen equation of state for the material was estimated using a rule of mixtures and existing shock and particle velocity data. Taylor cylinder tests were conducted and the rate of change in length was measured using high-speed video. Simulation of the Taylor tests was conducted using the developed strength and equation of state model and compared to the experimental results for model validation and comparison. Both the Johnson-Cook and Zerilli-Armstrong models resulted in less than 1% error of the Taylor cylinder results before material fracture. Further development of a fracture model for this material is recommended for use in high strain rate modeling applications
Artificial Intelligence, Real Risks: Understanding - And Mitigating - Vulnerabilities in the Military Use of AI
No full textArtificial Intelligence (AI) is becoming ubiquitous in daily life, and war is no exception to the trend. Given the role of AI and machine learning in strategic competition, it is critical that we understand both the risks introduced by these systems and their ability to create a strategic advantage. Here, we explore adversarial methods used to exploit vulnerabilities in AI models through a base example of target identification. We also discuss ways in which these risks can be mitigated. From this analysis, we conclude that humans must remain in the loop when operationalizing AI, and that we must continue to invest in and encourage the ethical use of AI