Pub Date : 2022-09-12DOI: 10.1177/15485129221105084
Colonel Eric D Trias
Modeling and simulation (M&S) continue to play a significant role in military training and operations. With the growth of cyberspace in support of military operations, it is only appropriate for M&S to follow a similar trajectory. In cybersecurity in particular, M&S can provide invaluable training for the cyber workforce, both for defensive and offensive operations, testing of network defenses and to support operational resiliency. Threats to military operations increase in complexity as adversaries develop their multi-domain capabilities to exploit information networks and mission systems. Adversaries are looking to physically attack defense critical infrastructure through cyber means along with exploiting vulnerability of information systems to gain physical access. Organizations must contend with threats from both physical and cyber means. A promising approach to assure operations resiliency in the face of this multi-domain threat lies in the concept of convergence of three security disciplines—physical, cyber, and continuity of operations (COOPs). Units can no longer depend on cybersecurity, nor can they rely entirely on guards, guns, and gates to protect critical missions, people, and infrastructure. Comprehensive risk-managed operational practices complemented by diverse, converged security protection programs are needed to meet these challenges. M&S has a significant role to play and must incorporate a more complex, holistic operational environment to address the resiliency of modern infrastructure. Network operators and cybersecurity providers must focus on assuring operational resilience and not merely on compliance with policies. Although policies provide a baseline to address common vulnerabilities, they are not sufficient in securing against complex threats, undiscovered vulnerabilities, or advanced adversaries. These adversaries continue to circumvent defenses whether from the inside, e.g., phishing and ransomware, or the outside through supply chain, vulnerable interfaces, or protocols. One way the Department of Defense (DoD) is addressing complex risks to its most strategic assets is to conduct a comprehensive vulnerability assessment utilizing a multidisciplinary approach called mission assurance (MA). MA, governed by DoD Instruction 3020.45, is the process to identify, protect, or ensure the continued function and resilience of capabilities and assets, including personnel, equipment, facilities, networks, information and information systems, infrastructure, and supply chains, critical to the execution of DoD mission-essential functions in any operating environment or condition. A major component of the MA concept is the on-site vulnerability assessment designed to discover gaps and weaknesses from multiple disciplines, i.e., physical security, general engineering, emergency management, and cyber operations. The framework provides a comprehensive risk assessment of critical assets that could prevent accomplishment of a unit, insta
{"title":"Modeling and simulation in mission assurance","authors":"Colonel Eric D Trias","doi":"10.1177/15485129221105084","DOIUrl":"https://doi.org/10.1177/15485129221105084","url":null,"abstract":"Modeling and simulation (M&S) continue to play a significant role in military training and operations. With the growth of cyberspace in support of military operations, it is only appropriate for M&S to follow a similar trajectory. In cybersecurity in particular, M&S can provide invaluable training for the cyber workforce, both for defensive and offensive operations, testing of network defenses and to support operational resiliency. Threats to military operations increase in complexity as adversaries develop their multi-domain capabilities to exploit information networks and mission systems. Adversaries are looking to physically attack defense critical infrastructure through cyber means along with exploiting vulnerability of information systems to gain physical access. Organizations must contend with threats from both physical and cyber means. A promising approach to assure operations resiliency in the face of this multi-domain threat lies in the concept of convergence of three security disciplines—physical, cyber, and continuity of operations (COOPs). Units can no longer depend on cybersecurity, nor can they rely entirely on guards, guns, and gates to protect critical missions, people, and infrastructure. Comprehensive risk-managed operational practices complemented by diverse, converged security protection programs are needed to meet these challenges. M&S has a significant role to play and must incorporate a more complex, holistic operational environment to address the resiliency of modern infrastructure. Network operators and cybersecurity providers must focus on assuring operational resilience and not merely on compliance with policies. Although policies provide a baseline to address common vulnerabilities, they are not sufficient in securing against complex threats, undiscovered vulnerabilities, or advanced adversaries. These adversaries continue to circumvent defenses whether from the inside, e.g., phishing and ransomware, or the outside through supply chain, vulnerable interfaces, or protocols. One way the Department of Defense (DoD) is addressing complex risks to its most strategic assets is to conduct a comprehensive vulnerability assessment utilizing a multidisciplinary approach called mission assurance (MA). MA, governed by DoD Instruction 3020.45, is the process to identify, protect, or ensure the continued function and resilience of capabilities and assets, including personnel, equipment, facilities, networks, information and information systems, infrastructure, and supply chains, critical to the execution of DoD mission-essential functions in any operating environment or condition. A major component of the MA concept is the on-site vulnerability assessment designed to discover gaps and weaknesses from multiple disciplines, i.e., physical security, general engineering, emergency management, and cyber operations. The framework provides a comprehensive risk assessment of critical assets that could prevent accomplishment of a unit, insta","PeriodicalId":44661,"journal":{"name":"Journal of Defense Modeling and Simulation-Applications Methodology Technology-JDMS","volume":null,"pages":null},"PeriodicalIF":0.8,"publicationDate":"2022-09-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"78037080","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-08-26DOI: 10.1177/15485129221118100
Phillip Pournelle
{"title":"The need for cooperation between wargaming and modeling & simulation for examining Cyber, Space, Electronic Warfare, and other topics","authors":"Phillip Pournelle","doi":"10.1177/15485129221118100","DOIUrl":"https://doi.org/10.1177/15485129221118100","url":null,"abstract":"","PeriodicalId":44661,"journal":{"name":"Journal of Defense Modeling and Simulation-Applications Methodology Technology-JDMS","volume":null,"pages":null},"PeriodicalIF":0.8,"publicationDate":"2022-08-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"89390298","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-08-05DOI: 10.1177/15485129221110893
Fairul Mohd-Zaid, Christine M. Schubert Kabban, R. Deckro, Wright Shamp
Social network analysis (SNA) is a tool for the operations researcher to understand, monitor, and exploit social and military structures which are key in the intelligence community. However, in order to study and influence a network of interest, the network must first be characterized; preferably to a known network model that captures a mixture of graphical properties exhibited by the social network of interest. In this work, we present a novel statistical method for both characterizing networks via a Binomial-Pareto maximum-likelihood approach and simulating the characterized network using a graph of mixed Barabási–Albert (BA, scale-free) and Erdös–Rényi (ER, randomness) properties. Characterization is performed through a combination of hypothesis tests and method of moments parameter estimation on Pareto and Doubly Truncated Binomial distributions. Application on real-world networks suggests that such networks may be characterized with a mixture of scale-free and random properties as modeled through BA and ER graphs. We demonstrate that our simulation methods are able to capture the degree distribution and density of the networks examined. These results demonstrate that this work establishes a statistical framework upon which network characterization and simulation may be accomplished, thus enabling the adaptation of such methods when generating, manipulating, and observing networks of interest.
{"title":"Network characterization and simulation via mixed properties of the Barabási–Albert and Erdös–Rényi degree distribution","authors":"Fairul Mohd-Zaid, Christine M. Schubert Kabban, R. Deckro, Wright Shamp","doi":"10.1177/15485129221110893","DOIUrl":"https://doi.org/10.1177/15485129221110893","url":null,"abstract":"Social network analysis (SNA) is a tool for the operations researcher to understand, monitor, and exploit social and military structures which are key in the intelligence community. However, in order to study and influence a network of interest, the network must first be characterized; preferably to a known network model that captures a mixture of graphical properties exhibited by the social network of interest. In this work, we present a novel statistical method for both characterizing networks via a Binomial-Pareto maximum-likelihood approach and simulating the characterized network using a graph of mixed Barabási–Albert (BA, scale-free) and Erdös–Rényi (ER, randomness) properties. Characterization is performed through a combination of hypothesis tests and method of moments parameter estimation on Pareto and Doubly Truncated Binomial distributions. Application on real-world networks suggests that such networks may be characterized with a mixture of scale-free and random properties as modeled through BA and ER graphs. We demonstrate that our simulation methods are able to capture the degree distribution and density of the networks examined. These results demonstrate that this work establishes a statistical framework upon which network characterization and simulation may be accomplished, thus enabling the adaptation of such methods when generating, manipulating, and observing networks of interest.","PeriodicalId":44661,"journal":{"name":"Journal of Defense Modeling and Simulation-Applications Methodology Technology-JDMS","volume":null,"pages":null},"PeriodicalIF":0.8,"publicationDate":"2022-08-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"78452870","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-08-05DOI: 10.1177/15485129221115740
R. Wallace
We explore the effects of Clausewitzian fog and friction using a data rate theorem–based model of the phase transition from control to failure for inherently unstable systems that include, but are not limited to, the many possible modalities of organized conflict. Fog-and-friction challenge any and all cognitive structures facing dynamic patterns of threat or opportunity, whether control is manifested through an institution, a machine entity, or some composite. The fundamental nature of challenge appears independent of the degree of sophistication of those institutions, entities, or composites, and of the technical modalities employed. The dialog/Zweikampf of organized conflict is—and will remain—an intimate and most human enterprise. Implications for other existential threats of inherently unstable circumstance, like pandemic disease or climate change, are evident.
{"title":"Fog, friction, and control in organized conflict: punctuated transitions to instability","authors":"R. Wallace","doi":"10.1177/15485129221115740","DOIUrl":"https://doi.org/10.1177/15485129221115740","url":null,"abstract":"We explore the effects of Clausewitzian fog and friction using a data rate theorem–based model of the phase transition from control to failure for inherently unstable systems that include, but are not limited to, the many possible modalities of organized conflict. Fog-and-friction challenge any and all cognitive structures facing dynamic patterns of threat or opportunity, whether control is manifested through an institution, a machine entity, or some composite. The fundamental nature of challenge appears independent of the degree of sophistication of those institutions, entities, or composites, and of the technical modalities employed. The dialog/Zweikampf of organized conflict is—and will remain—an intimate and most human enterprise. Implications for other existential threats of inherently unstable circumstance, like pandemic disease or climate change, are evident.","PeriodicalId":44661,"journal":{"name":"Journal of Defense Modeling and Simulation-Applications Methodology Technology-JDMS","volume":null,"pages":null},"PeriodicalIF":0.8,"publicationDate":"2022-08-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"80127573","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-08-05DOI: 10.1177/15485129221115736
Maarten P. D. Schadd, Anne Merel Sternheim, R. Blankendaal, Martin van der Kaaij, Olaf H. Visker
With recent technological advances, commanders request the support of artificial intelligence (AI)-enabled systems during mission planning. Future AI systems may test a wide range of courses of action (COAs) and use a simulator to test each COA’s effectiveness in a war game. The COA’s effectiveness is however dependent on the commanders’ intent. The question arises to what degree a machine can understand the commanders’ intent? Currently, the intent has to be programmed manually, costing valuable time. Therefore, we tested whether a tool can understand a freely written intent so that a commander can work with an AI system with minimal effort. The work consisted of letting a tool understand the language and grammar of the commander to find relevant information in the intent; creating a (visual) representation of the intent to the commander (back brief); and creating an intent-based computable measure of effectiveness. We proposed a novel quantitative evaluation metric for understanding the commanders’ intent and tested the results qualitatively with platoon commanders of the 11th Airmobile Brigade. They were positively surprised with the level of understanding and appreciated the validation feedback. The computable measure of effectiveness is the first step toward bridging the gap between the command intent and machine learning for military mission planning.
{"title":"How a machine can understand the command intent","authors":"Maarten P. D. Schadd, Anne Merel Sternheim, R. Blankendaal, Martin van der Kaaij, Olaf H. Visker","doi":"10.1177/15485129221115736","DOIUrl":"https://doi.org/10.1177/15485129221115736","url":null,"abstract":"With recent technological advances, commanders request the support of artificial intelligence (AI)-enabled systems during mission planning. Future AI systems may test a wide range of courses of action (COAs) and use a simulator to test each COA’s effectiveness in a war game. The COA’s effectiveness is however dependent on the commanders’ intent. The question arises to what degree a machine can understand the commanders’ intent? Currently, the intent has to be programmed manually, costing valuable time. Therefore, we tested whether a tool can understand a freely written intent so that a commander can work with an AI system with minimal effort. The work consisted of letting a tool understand the language and grammar of the commander to find relevant information in the intent; creating a (visual) representation of the intent to the commander (back brief); and creating an intent-based computable measure of effectiveness. We proposed a novel quantitative evaluation metric for understanding the commanders’ intent and tested the results qualitatively with platoon commanders of the 11th Airmobile Brigade. They were positively surprised with the level of understanding and appreciated the validation feedback. The computable measure of effectiveness is the first step toward bridging the gap between the command intent and machine learning for military mission planning.","PeriodicalId":44661,"journal":{"name":"Journal of Defense Modeling and Simulation-Applications Methodology Technology-JDMS","volume":null,"pages":null},"PeriodicalIF":0.8,"publicationDate":"2022-08-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"72657265","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-08-05DOI: 10.1177/15485129221111171
Melissa Pescatore, Paul T. Beery
This paper demonstrates an approach for the use of agent-based simulation, supported by model-based systems engineering products, to analyze interoperability. To demonstrate the approach, a representative maritime search-and-rescue (SAR) operation is simulated in the agent-based simulation program Map-Aware Non-Uniform Automata (MANA). The MANA SAR model is used to assess interoperability decisions at organizational, operational, and technical levels and to highlight dependencies between decisions at each level of interoperability. Analysis indicates that, within the MANA SAR model, organizational interoperability decisions have the largest impact on operational performance but that organizational challenges may be overcome with substantial investment at both the operational and technical levels of interoperability.
{"title":"Interoperability analysis via agent-based simulation","authors":"Melissa Pescatore, Paul T. Beery","doi":"10.1177/15485129221111171","DOIUrl":"https://doi.org/10.1177/15485129221111171","url":null,"abstract":"This paper demonstrates an approach for the use of agent-based simulation, supported by model-based systems engineering products, to analyze interoperability. To demonstrate the approach, a representative maritime search-and-rescue (SAR) operation is simulated in the agent-based simulation program Map-Aware Non-Uniform Automata (MANA). The MANA SAR model is used to assess interoperability decisions at organizational, operational, and technical levels and to highlight dependencies between decisions at each level of interoperability. Analysis indicates that, within the MANA SAR model, organizational interoperability decisions have the largest impact on operational performance but that organizational challenges may be overcome with substantial investment at both the operational and technical levels of interoperability.","PeriodicalId":44661,"journal":{"name":"Journal of Defense Modeling and Simulation-Applications Methodology Technology-JDMS","volume":null,"pages":null},"PeriodicalIF":0.8,"publicationDate":"2022-08-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"84405022","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
D. Tarraf, J. Gilmore, D. Barnett, Scott S. Boston, David Frelinger, Daniel C. Gonzales, Alexander C. Hou, Peter Whitehead
In this report, researchers experimented with how postulated artificial intelligence/machine learning (AI/ML) capabilities could be incorporated into a wargame. We modified and augmented the rules and engagement statistics used in a commercial tabletop wargame to enable (1) remotely operated and fully autonomous combat vehicles and (2) vehicles with AI/ML-enabled situational awareness to show how the two types of vehicles would perform in company-level engagement between Blue (US) and Red (Russian) forces. The augmented rules and statistics we developed for this wargame were based in part on the US Army’s evolving plans for developing and fielding robotic and AI/ML-enabled weapon and other systems. However, we also portrayed combat vehicles with the capability to autonomously detect, identify, and engage targets without human intervention, which the Army does not presently envision. The rules we developed sought to realistically portray the capabilities and limitations of AI/ML-enabled systems, including their vulnerability to selected enemy countermeasures, such as jamming. Future work could improve the realism of both the gameplay and representation of AI/ML-enabled systems, thereby providing useful information to the acquisition and operational communities in the US Department of Defense.
{"title":"An Experiment in Tactical Wargaming with Platforms Enabled by Artificial Intelligence","authors":"D. Tarraf, J. Gilmore, D. Barnett, Scott S. Boston, David Frelinger, Daniel C. Gonzales, Alexander C. Hou, Peter Whitehead","doi":"10.7249/rra423-1","DOIUrl":"https://doi.org/10.7249/rra423-1","url":null,"abstract":"In this report, researchers experimented with how postulated artificial intelligence/machine learning (AI/ML) capabilities could be incorporated into a wargame. We modified and augmented the rules and engagement statistics used in a commercial tabletop wargame to enable (1) remotely operated and fully autonomous combat vehicles and (2) vehicles with AI/ML-enabled situational awareness to show how the two types of vehicles would perform in company-level engagement between Blue (US) and Red (Russian) forces. The augmented rules and statistics we developed for this wargame were based in part on the US Army’s evolving plans for developing and fielding robotic and AI/ML-enabled weapon and other systems. However, we also portrayed combat vehicles with the capability to autonomously detect, identify, and engage targets without human intervention, which the Army does not presently envision. The rules we developed sought to realistically portray the capabilities and limitations of AI/ML-enabled systems, including their vulnerability to selected enemy countermeasures, such as jamming. Future work could improve the realism of both the gameplay and representation of AI/ML-enabled systems, thereby providing useful information to the acquisition and operational communities in the US Department of Defense.","PeriodicalId":44661,"journal":{"name":"Journal of Defense Modeling and Simulation-Applications Methodology Technology-JDMS","volume":null,"pages":null},"PeriodicalIF":0.8,"publicationDate":"2022-05-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"89728564","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-01DOI: 10.1177/15485129221096478
James Ryseff, M. Bond
{"title":"Small is beautiful","authors":"James Ryseff, M. Bond","doi":"10.1177/15485129221096478","DOIUrl":"https://doi.org/10.1177/15485129221096478","url":null,"abstract":"","PeriodicalId":44661,"journal":{"name":"Journal of Defense Modeling and Simulation-Applications Methodology Technology-JDMS","volume":null,"pages":null},"PeriodicalIF":0.8,"publicationDate":"2022-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"75946416","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-03-25DOI: 10.1177/15485129221088717
M. van Lent, D. Schmorrow
{"title":"The applications of artificial intelligence to education and training","authors":"M. van Lent, D. Schmorrow","doi":"10.1177/15485129221088717","DOIUrl":"https://doi.org/10.1177/15485129221088717","url":null,"abstract":"","PeriodicalId":44661,"journal":{"name":"Journal of Defense Modeling and Simulation-Applications Methodology Technology-JDMS","volume":null,"pages":null},"PeriodicalIF":0.8,"publicationDate":"2022-03-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"79298449","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-03-19DOI: 10.1177/15485129221088718
S. Schatz, J. Walcutt
Let’s be honest, artificial intelligence (AI) will change— or, rather, is already changing—so much. It would be easy, if uninspired, to fill this article with a laundry list. But rather than add to the existing litany of forecasts (many of which you can read in the chapters of this special edition), we’ll focus more narrowly. First, we’ve bound the question to learning in the defense domain, and second, we’ve challenged ourselves to target a single concept—to name the linchpin with greatest potential to have profound, paradigm-changing impacts. To give away the punchline, we’ve selected ‘‘the way we measure and evaluate.’’ Before we show our work, consider these definitions. Measure and evaluate refer to two sides of the same coin. Formally, measurement is the ‘‘quantitatively expressed reduction of uncertainty based on one or more observations’’ (p. 23). In other words, it refers to collected observations (no matter how fuzzy or incomplete) that help us fill-in (but not necessarily eliminate) uncertainty in a Claude Shannon ‘‘information theory’’ sort of way. Measurement goes hand-in-hand with evaluation. Evaluation is the process of interpreting the data collected from measurements, and for our purposes, we’ll say it covers all of the associated aggregation, transformation, analysis, and other activities needed to effectively use the measured data. Learning, as a formal concept, is related to—but notably distinct from—training and education. Those latter two terms, particularly in a defense context, are laden with connotations. ‘‘Training and education’’ refer to the organizational side of the experience, for instance, to the curriculum or the wargame delivered by a schoolhouse or training branch. They’re input-focused terms, and more than that, they tend to imply a formal learning context. In contrast, the term ‘‘learning’’ focuses on the individual (or team) side of the equation—the outcomes side. It describes any change in long-term memory that affects knowledge, skills, or behaviors, and it makes no distinction for the process through which it was acquired. 1. An operational perspective
{"title":"Modeling what matters: AI and the future of defense learning","authors":"S. Schatz, J. Walcutt","doi":"10.1177/15485129221088718","DOIUrl":"https://doi.org/10.1177/15485129221088718","url":null,"abstract":"Let’s be honest, artificial intelligence (AI) will change— or, rather, is already changing—so much. It would be easy, if uninspired, to fill this article with a laundry list. But rather than add to the existing litany of forecasts (many of which you can read in the chapters of this special edition), we’ll focus more narrowly. First, we’ve bound the question to learning in the defense domain, and second, we’ve challenged ourselves to target a single concept—to name the linchpin with greatest potential to have profound, paradigm-changing impacts. To give away the punchline, we’ve selected ‘‘the way we measure and evaluate.’’ Before we show our work, consider these definitions. Measure and evaluate refer to two sides of the same coin. Formally, measurement is the ‘‘quantitatively expressed reduction of uncertainty based on one or more observations’’ (p. 23). In other words, it refers to collected observations (no matter how fuzzy or incomplete) that help us fill-in (but not necessarily eliminate) uncertainty in a Claude Shannon ‘‘information theory’’ sort of way. Measurement goes hand-in-hand with evaluation. Evaluation is the process of interpreting the data collected from measurements, and for our purposes, we’ll say it covers all of the associated aggregation, transformation, analysis, and other activities needed to effectively use the measured data. Learning, as a formal concept, is related to—but notably distinct from—training and education. Those latter two terms, particularly in a defense context, are laden with connotations. ‘‘Training and education’’ refer to the organizational side of the experience, for instance, to the curriculum or the wargame delivered by a schoolhouse or training branch. They’re input-focused terms, and more than that, they tend to imply a formal learning context. In contrast, the term ‘‘learning’’ focuses on the individual (or team) side of the equation—the outcomes side. It describes any change in long-term memory that affects knowledge, skills, or behaviors, and it makes no distinction for the process through which it was acquired. 1. An operational perspective","PeriodicalId":44661,"journal":{"name":"Journal of Defense Modeling and Simulation-Applications Methodology Technology-JDMS","volume":null,"pages":null},"PeriodicalIF":0.8,"publicationDate":"2022-03-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"82517201","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：