A review of applications in Artificial Intelligence (AI) on the Security and defense field

L'article suivant a été fourni par une source externe. Le gouvernement du Canada et le Collège militaire royal de Saint-Jean n'assument aucune responsabilité concernant la précision, l'actualité ou la fiabilité des informations fournies par les sources externes. Les utilisateurs qui désirent employer cette information devraient consulter directement la source des informations. Le contenu fourni par les sources externes n'est pas assujetti aux exigences sur les langues officielles, la protection des renseignements personnels et l'accessibilité.

Fabián Garay-Rairán, ESJIM Research Group, Non-Commissioned Officers School Colombian Police, Sibaté, Colombia; Research Group of the Center of Simulation and Research (CISI), Infantry School ESINF, Bogotá, Colombia

Miguel Bermúdez-Melo, Colombian National Army; Research Group of the Center of Simulation and Research (CISI), Infantry School ESINF, Bogotá, Colombia

Hernando Guerrero-Gómez, Research Group of the Center of Simulation and Research (CISI), Infantry School ESINF, Bogotá, Colombia

Carlos Peña-Lucumí, ESJIM Research Group, Non-Commissioned Officers School Colombian Police



The branch or field of computer science that develops processes that imitate the intelligence of living beings is called Artificial Intelligence (AI) (Hardy, 2001). There have been many advances in recent history to achieve the reproduction of the cognitive capacities of the human being in machines. AI is in charge of channeling such research efforts in this regard.

When speaking about AI we must consider subjects such as: development, innovation, and new technologies, aspects that are not at all unrelated to the military environment, which leads to the following question - what would happen if these two fields were joined? It hopefully would not be counterproductive and would manage to generate new technologies to help solve problems related to communications, health, and other security and defense issues. Innovation and the updating of conventional defense and security methods is of vital importance for military and police realms, since potential adversaries on all fronts are redefining conflict and using appropriate weapons for dominance (Allenby, 2016).

Due to the aforementioned, there is a need to review the applications of AI technologies for security and defense environments, clarifying the different branches of where it can be applied as well as the benefits that innovative AI applications can bring. This paper will conduct an extensive review of AI technologies considering sources from scientific bibliographies, data bases and specialized journals. It aims to reveal new approaches, tendencies, and applications that make use of AI inside technological innovations by different armies around the world.


Different methods were used to locate the bibliographic documents. A bibliographic search was carried out in the IEEE Xplore database, for which the descriptors used were: artificial intelligence, military, army, police, neural networks, health, communications, simulation, and vehicles. The results obtained by subject area were around 23 and 12 after the combination of the different key words. An Internet search was also carried out in Google Scholar and in specialized databases such as Embase, Springer Link, Science Direct, and Scopus; the same search terms were used. Documents that included the construction of prototypes with military applications based on AI were selected.

The selection of documents was made based on an analysis of the reliability of the results, the relevance of their findings, and their applicability to the area of study. Further, the information was organized in consideration of the application or development of the technologies on AI aimed at improving processes concerning security and defense within five main areas: 1) training and simulation, 2) health and medicine, 3) telecommunications, 4) military vehicles and 5) applications for police departments. Then, the information under consideration was organized and the PRISMA statement was used for systematic document review, making use of the verification recommendations, regarding the preparation of structured summaries. Finally, through the RStudio Cloud software, the Scopus queries were analyzed, resulting in the most important documents by author (the aforementioned search terms were used).


A fundamental part of military training rests on its strict individual training (conventional soldiering skills). However, over the years traditional military training has begun to become obsolete when competing with the innovation of new trends in military training. This includes the use of simulators for weapon systems, vehicles, ranges, and other applications (Pereira Rueda, 2011). In reality, new technologies permit many novel ways to approach training, this includes the use of AI. This includes, simulation which is postulated as a great advantage, reducing costs in unnecessary losses and generating high-caliber knowledge without the need to expose users to danger when being trained.

For instance, there are games focused on military training that already have a long history of application, such as in the form of sand boxes, mockups, war games and computer games. The dynamic representations of the physical world in games allow a more immersive learning experience for Army personnel, employing some of the most advanced technologies in computing (Smith, 2010). Among them the use of AI clearly stands out because of its ability to program opponents with great detail. Simulation games for military training offer the following benefits:

  • A 3D engine to visualize the surroundings in clear and realistic way
  • User-friendly graphical interface and interactive patterns that allow the player to start without reading the manual
  • Physical models maintain the notion of reality and also help to maintaining the basic parameters of movement, such as gravity, trajectories, and collisions
  • Using AI to program opposing forces to increase the difficulty of the training and in turn creating a different interaction in each attempt

Using AI provides the computational power necessary to “create in-game opponents that are smart enough to challenge human players”, creating an adaptive experience for the user that can adjust with game progression (Smith, 2010, p. 7). In this way, AI is implemented in the world of "serious games" (Game designed for a purpose other than entertainment) to improve the mechanics of military training (Smith, 2010). Since 1994, the discussion of using AI has been on the table, to solve the complexity of increasingly large and interactive simulations that are difficult to administer. In this case AI and artificial neural networks (ANN) facilitated mechanisms to automate command and control, through linear and interactive models of high-speed object-oriented activation, which facilitates its use in various battle simulations (Jaszlics et al., 1994).

In the US Army military simulation applications come into action, more specifically agent-based artificial life simulations. This includes in an almost intuitive way the Defense Department, who uses simulations for their decision-making process (Cioppa et al, 2004). ABS (Agent-based Simulation), is an approach to modeling systems comprised of individual, autonomous, interacting “agents” widely used for military simulations, which have already proven their usefulness to generate ideas and focus simulation experiments., Utility was confirmed in an experiment carried out in the United States where an attack situation was presented to a combined arms battalion. Three different types of ABS were put to the test for their effectiveness and performance, resulting in the victory of the Map Aware Non-Uniform Automata (MANA). The MANA model “is an agent-based model, a particular class of models in which entities (called agents) have simple rules defining their behavior” (Wilson, 2017, p.6). MANA facilitates quick construction and exploration of new scenarios, the GUI helps to create scenarios and the playback functions are useful analysis tools that “allows considerable flexibility in creating a diverse set of agents” (Cioppa et al., 2004, p.174).

The MANA model maintains a “memory of the battlefield” and their behaviors can adapt to a variety of battlefield events (Cioppa et al., 2004, p. 174). Consequently, in this example, the effectiveness of AI as an agent is seen. Work is still being done to improve and harness the ability of ABS to aid military decision-making processes. In addition, these AI applications are not only used in battle simulations but can be used to analyze technical, commercial, and even infrastructural problems.

The advancement of technology has led to the battlefield to move to other more virtual universes and it is there where cyber-defense plays a leading role; therefore, intelligent and autonomous AI agents will be widely present in the battlefield of the future and of the present as well, which will likewise be a greater part of the total military assets. Therefore, in future cyber defence, autonomous and intelligent agents will become the main cyber combatants in the future battlefield (Kott, 2018).


For various reasons, the health of Armed Forces personnel has always been a delicate subject. This ranges from institution’s internal requirements for membership in relation to the physical state and health of the body, to cases of poor hygiene and sanitation in medical supplies and first aid kits, along with a myriad of other health related topics. This is where new technologies, more specifically AI come to play an important role.

Physiological tests in the military are very common and the traditional way of doing them has become tedious and a bit outdated, since it can involve human errors with or without intention. The solution to this is relatively simple and consists of putting these tasks in the hands of technology. US researchers have developed a device to detect, receive, obtain and present human physiological and contextual information that includes a series of sensors to collect data related to a physiological state and various contextual parameters of an individual (Stivoric et al., 2007).

To this point AI has already replaced the conventional way of taking physiological examinations and offers significant improvements. Examination devices can be programmed based on AI that modifies the exercise program followed by the user based on the information collected. This algorithm development process can create algorithms to enable the device to detect and measure various parameters, including without limitation, the following:

  • When an individual is experiencing a seizure, states of unconsciousness, fatigue, shock, drowsiness, heat stress, and dehydration
  • A state of the individual's availability, health, and / or metabolic state, such as in a military setting, which includes the states of dehydration, malnutrition, and lack of sleep

AI applications can eliminate clinical uncertainty that may lead to seemingly contradictory conclusions, as the above invention weighs important joint probabilities and determines the most likely conclusion (Stivoric et al., 2017).

Other applications of AI in the field of health prevention have applied machine learning and neural networks. One application involves forecasting the growth of microcystin toxin levels above US Environmental Protection Agency (EPA) health limits. In a US Army Corps of Engineers reservoir, a harmful algae bloom (HAB) was predicted one week in advance. As a result, a potential for predicting toxicity (microcystin ELISA) intervals of 95 percent was obtained and allowing other probabilistic prognosis products (Fleming et al., 2019).

In the case of the Colombian National Army, the Infantry School (Garay, et al, 2019) carried out a project to improve the quality of life of infantry personnel through the use of AI. Colombian Army personnel are exposed to long periods of training and work, which involve abrupt, sudden, and unexpected movements of the body, together with excessive loads on their backs, which generates spinal disorders in the short, medium and long term. This, affects the quality of life of the soldier and has a great economic impact on their families, the Army, and the Colombian Government due to absenteeism, incapacity for work and cost of specialized medical care (Garay et al., 2019). This work presented and compared various training methods for a computer-assisted system capable of evaluating and classifying the health status of the spine amongst military personnel. This system was validated by evaluating its computational performance and by diagnostics of utility measurements which demonstrated that the simple genetic algorithm is the best solution to the problem and is an excellent tool to help military health professionals reduce diagnostic uncertainty (Garay et al., 2019). Figure 1, illustrates the structure of the fuzzy system implemented for the application of the Infantry School.

Fig 1. Architecture of the Mamdani- type fuzzy system proposed for the evaluation of the proposed classification systems. Taken from: (Garay et al., 2019).

The possibilities of AI in the field of radiology for medical services of the military have been described in several studies, especially the possibility of improving processes or work that is by nature tedious, repetitive and slow, in the detection within diagnostic images. In terms of military applications, AI can be incorporated into radiological services assisting small hospitals and remote detachments without radiologists to alleviate the burden on larger hospitals and to avoid the movement of military personnel to more specialized hospitals (Sen et al., 2019).

Finally, AI has become popular as a military application for the prediction of suicide risk, a public health phenomenon with a worldwide prevalence of approximately 800,000 deaths per year. AI applications will continue to assist clinical management and mental health care (Fonseka et al., 2019). AI has emerged as a promising suicide reduction tool that would complement and improve the results of prediction models carried out within the US Army. The study developed by Bernecker et al., (2019) used a number of smaller variables in a self-administered survey within a representative sample of the population to create a prediction model which found that 10 percent of the sample identified as high risk. Previous research such as Kessler et al., (2017) had already concluded that an even better model could be developed with enriched information on the risk of suicide assessed by the doctor enhanced with the use of other tools.


AI techniques can also improve the robustness and tolerance of error within communication systems (Schutzer, 1983). The control and maintenance of communication systems is generally an intensive process that requires qualified experts with extensive training and experience. “Expert system technology can be used to aid in the control and maintenance of a communication system” (Schutzer, 1983, p.789). In general, “these experts consciously or unconsciously apply rules of thumb to diagnose and to detect abnormal and faulty communications” (Schutzer, p.789). Recommended corrective actions are also derived from these collections of general heuristic rules. Expert systems technologies, such as production rule systems, allow these rules to be acquired and represented in a computer-compatible way (Schutzer, 1983). The aforementioned was clearly evidenced in a project carried out by the US Navy in which it was demonstrated how natural processing languages and knowledge-based systems technologies could improve quality of communication systems (Schutzer, 1983).

Schutzer (1983) suggests that protocols of vocabulary, grammar, and common context knowledge should be agreed upon in advance to facilitate communication using short code words. Distributed databases must also be updated and kept up to date. AI techniques also help reduce the associated communications overhead for this upgrade process. The results of the research were positive, but clearly, the applicability of this technique depends a lot on the situation, cases like the following:

  • “The more dynamic, uncertain, and unpredictable an object’s future behavior, the more frequent are the required data updates and the less that is gained through generation of memory-dependent data base and inference mechanisms” (Schutzer, 1983, p. 789).
  • “When the data updates are very noisy and unreliable, the need for a memory-dependent database is justified, not only to save communication overhead, but also to increase the quality and reliability of the inferred tactical situation” (Schutzer, 1983, p. 789).
  • “Received sensor detections can be validated and checked for consistency and logical plausibility” (Schutzer, 1983, p. 789).

Previous observations can be included with built-in rules, pre-determined values, and procedures to infer probable future outcomes (Schutzer, 1983). In addition to improving the robustness and tolerance of communication systems, AI can also be used to support noise monitoring stations. These monitoring stations allow to assess in real-time the impact inside and outside of a process or operation, noise monitoring stations involve several main elements where they stand out:

  • Sound level meter with microphone for this type of measurements
  • Assembly and installation of infrastructure
  • Data transmission system

Bucci and Vipperman (2007) used noise monitoring stations located around some military installations to assist in the processing of noise claims and damage claims, but despite their proper installation and operation they can produce false positives and may not correctly detect impulse events. To avoid this, classifiers using artificial neural networks was developed (Bucci & Vipperman, 2007) to improve the precision of the identification of military impulse noise. For this, a neural network algorithm was created that received two time domain indicators:

  • Kurtosis
  • Crest factor

And two custom frequency domain metrics:

  • Spectral slope
  • Weighted square error

Algorithm development included a training and an evaluation phase (Bucci & Vipperman, 2007). The classification algorithm was able to achieve up to 100 percent accuracy in training data and validation data, while improving the detection threshold by at least 40 dB. All the algorithms were developed in MATLAB and could be applied in real time (Bucci & Vipperman, 2007). In this way, they guarantee efficiency and accuracy on the different types of noise pollution, classifying each of them in its respective group, to find detectable failures by hearing index and avoid unjustified claims.

Other applications in the communications sector have made use of AI for data reduction in the use of data in applications of key management areas in the US Army. This application is necessary as the US Army began the development of a work operator station to provide an integrated software package for COMSEC information processing, signal operating instructions, electronic countermeasures, and network communication planning. The use of AI allows taking advantage of the systems and to better their response (Pinsky & Lynn, 1992).


Across both land and air platforms It is normal to find errors related to the war functions (maneuvering and movement) of vehicles. For example, there can be errors related to the registration of information, registration relative to maintenance, inspection and technical support of the vehicles. Errors include:

  • Loss of information.
  • Erroneous recording of information.
  • The standardization processes are not carried out
  • The late delivery of the vehicle information
  • Lack of standardized processes for obtaining specific information in real time on the current status of vehicles in operation or maintenance.
  • False alarms due to human errors.

These types of problems can be solved with the use of systematic tools that use AI for data acquisition and reporting. There is a project in the US that describes the development of a prototype for an “integrated diagnostic system to optimize maintenance on military transport aircraft by reducing maintenance costs and increasing aircraft availability” (Chidambaram et al., 2006, p. 1). This project addressed three specific maintenance functions, identifying the root cause of an error message, identify false alarms, planning maintenance. In addition, the prototype system comprises an on-board and a ground-based component, solid-state high-capacity data logger and an Advanced Wireless Open Data Systems (AWODS) that monitors and records data from the bus avionics during flight. The data is downloaded, at the end of the flight, for analysis by the ground component. There are several utilities of this approach, including root cause identification, false alarm filtering, and prioritization of maintenance actions. These systems improve diagnostic aircraft data (Chidambaram et al., 2006).

Technologies used to analyze the data included: expert systems / rapid state recognition methods, data mining and Bayesian analysis and was developed using OSA-CBM – “an open software standard for the development of diagnostic software” (Chidambaram et al., 2006, p.1). By emphasizing this integrated diagnostics and fault analysis mode, human failures and false alarms can be avoided thanks to the reasoning capabilities of the software due to Bayesian networks. In the field of vehicles, integrated diagnosis and estimation of maintenance time are applied for complex engineering systems, with AI systems that allow predicting the “future health status of a system or component, as well as providing the ability to anticipate faults, problems, potential failures and required maintenance actions” (Azam et al., 2014, p.1). This facilitates the preparation for inspection actions and timely maintenance scheduling, minimizing downtime and optimizing costs (Azam et al., 2014).

Regarding autonomous vehicles, it is common in the military field to provide data to train learning agents. In this regard, research such as those conducted by Knox & Stone (2009) claim that “an autonomous driving agent should not learn to drive by crashing into road barriers and endangering the lives of pedestrians” (p.1). For this reason, the labels of the data with which an algorithm is trained are of the utmost importance. AI applications for aircraft make use of loggers and reasoners for integrated diagnosis which aim to optimize the maintenance of military transport aircraft, to decrease maintenance costs, and increase availability of aircraft (Chidambaram et al., 2006).


Police AI applications boomed specially in the use of algorithms for the analysis of crimes, in the area of home invasion, criminal profiling, criminal tracking, identifying people, and suspicious areas (Joh, 2017). Little by little AI has played an important role as a tool to support police operations. The Artificial Intelligence Crime Analysis and Management System (AICAMS) is a system that was born from a simple but effective methodology that evolved from the experience of several with projects developed by the Ottawa police, Hong Kong police and home invasion data from North America. This system based on rules makes use of the automatic learning and techniques of neural networks (Brahan et al., 1998). Figure 2 shows the block diagram of the AICAMS system.

Fig 2. Operation of AICAMS. Taken from: (Brahan et al., 1998).

Another AI software called Crime Similarity System (CSS) enables and facilitates police departments to develop a strategic point of view towards decision making through the use of socio-economic, criminal, and law enforcement data to identify populations who are more likely to share information. This allowed discovering opportunities for police departments to cooperate, improve their investigative capacity, and contribute to making the investigation permanently available in electronic media (Redmond & Baveja, 2002).

Another example is Rahul Shah’s project from California State University, whose objective was to develop a crime prediction system using machine learning (Rahul Shah, 2020). This was a strategy for determining crime regions and thus executing activities for prevention and mitigation of crime rates. This application of AI makes use of the identification of patterns which will allow to address problems with unique approaches in regions of specific crime categories and improve more security measures in society. Other research by Bharati and Rak (2018) extracted crime data from the official website of the Chicago police consisting of the description of the location, type of offense, date, time, and latitude/longitude, the model used was the K- Nearest Neighbor (KNN) classification and other algorithms for crime prediction and crime solving at a much faster rate and thus reduced the crime rate.

Another application of AI was developed at the Indian Institute of Information Technology and Management-Kerala (IIITM-K), with the hope to predict and prevent future crime. This application was based on the application of the Auto Regressive Integrated Moving Average (ARIMA). The modeling involved dividing training data from1953 to 2008 and the test data for the years 2009 to 2013, as a result the model achieves a forecast value within the 95 percent confidence interval of the test data (Kumar et al., 2018).

Finally, it is worth highlighting the approach of Mcclendon & Meghanathan (2015), which used data mining and machine learning to carry out a comparative study between violent crime patterns from a community data set provided by the University of California-Irvine repository and actual crime statistics for the state of Mississippi.

It is important to acknowledge the limitations of knowledge-based expert systems. The technology can provide an organized approach to acquire pertinent information and help the investigator save time, however, they do not independently solve crimes (Brahan et al., 1998).


As early as 1988, the potential applications of expert systems within the defense sector have been discussed. Research such as Shah and Buckner (1988) carried out a review of the uses of expert systems and how they can affect the development of future defense applications, which at the time were already evidenced by the artificial research and development efforts of the military, especially in developments by NASA, the United States Air Force and in general in North American Strategic Defense initiatives. Today there is talk of new battle scenarios, so many armies are deploying multi-layered systems to fight in various domains, including space, cyber, air, maritime and land. It is here where AI, machine learning and expert systems will play a fundamental role as engines of change in the way in which conflicts unfold and assist those who will emerge victorious (Stanton et al., 2020).

Multiple research describes the transformation that war scenarios will have, such as in Bangladesh where the discussion about the role that machines and AI will play in future conflicts is already sitting. In this debate, Ahmed’s research (2019) mentions that appropriate measures must be taken to be able to face such technological hype in future war (Ahmed, 2019). The United States knows first-hand the value of innovating and being at the forefront of technological development and always very hand in hand with the technological dimension in the defense sector and currently the US Department of Defense, strives to apply AI to the army, in the following purposes: situational awareness, cybernetics, military logistics, command and control, and technology and swarm tactics for autonomous unmanned systems (Mori, 2018). The latter has been facilitated by the Institute of Innovation AI Center (A2I2) tasked with accelerating basic research to address specific Army challenges, through advancing AI capabilities for autonomous maneuvers in multidomain operations (MDO) (Cirincione et al., 2019; Stanton et al., 2020).

Most of the information available on AI and its military applications is concentrated in the US as the protagonist and naturally their analyses have laid the lines of current trends in AI and machine learning, within which they indicate the correct treatment of data since it is evident that large amounts of data is required for successful algorithm training (Bean, 2017). In military and police contexts, however, there is usually a large amount of data collected but it is not labeled or correctly labeled (Fossaceca & Young, 2018). The research areas that should be supported in the future for the transformation of the security and defense sector can be seen in Figure 3.

Fig 3. The 9 essential areas of research in the security and defense sector of the future. Adapted from: (Fossaceca & Young, 2018).

Today, a wide variety of algorithms are easy to find in the security and defence sector. Research facilitates new approaches to continually improving business methods for AI and machine learning, opening a window for these to be adapted and applied to priority scenarios. Finally, there are many challenges on the horizon, so priorities should also focus on better methods to provide “robust performance with noisy data, quantification of uncertainty” and treatment of attacks against data (cyberattacks) (Fossaceca & Young, 2018, p.16). This includes, the integration of the Internet of Battlefield Things (IoBT) and Collaborative Technology Alliance for the implementation of intelligent command and control and battlefield services (Fossaceca & Young, 2018, p.16).


AI offers various opportunities for improvement for militaries and police departments, such as the Colombian ones, who at this moment are undergoing very important transformational processes which in turn empower the technological modernization that is required. It is possible to direct future research and developments through implementations with AI, in the fields described above such as: training and simulation, medicine and health sciences, communications, military vehicles and police applications. Technological development has almost infinite possibilities that must begin to be discovered and exploited.

It is important that the technological surveillance centers of the security and defense sector have their eyes set on the advances in AI, expert systems and machine learning, since their applications are quite promising for this sector. However, as it was seen previously, practically the exploration of AI is relatively recent in the military and police fields, so its development will be increased in the short and medium term.


Ahmed, L. (2019). Emergence of Artificial Intelligence in future warfare: Preparedness of Bangladesh Armed Forces. NDC Journal, 179–209. https://ndcjournal.ndc.gov.bd/ndcj/index.php/ndcj/article/download/258/233

Allenby, B. (2016). What Exactly Is “Military Artificial Intelligence”? SLATE. http://www.slate.com/articles/technology/future_tense/

Azam, M., Ghoshal, S., Deb, S., Pattipati, K., Haste, D., Mandal, S., & Kleinman, D. (2014). Integrated diagnostics and time to maintenance estimation for complex engineering systems. 2014 IEEE Aerospace Conference, 1–10. https://doi.org/10.1109/AERO.2014.6836478

Bean, R. (2017, May 8). How Big Data Is Empowering AI and Machine Learning at Scale. Sloan Management Review. https://sloanreview.mit.edu/article/how-big-data-is-empowering-ai-and-machine-learning-at-scale/

Bernecker, S. L., Zuromski, K. L., Gutierrez, P. M., Joiner, T. E., King, A. J., Liu, H., Nock, M. K., Sampson, N. A., Zaslavsky, A. M., Stein, M. B., Ursano, R. J., & Kessler, R. C. (2019). Predicting suicide attempts among soldiers who deny suicidal ideation in the Army Study to Assess Risk and Resilience in Servicemembers (Army STARRS). Behaviour Research and Therapy, 120. https://doi.org/10.1016/j.brat.2018.11.018

Bharati, A., & Rak, S. (2018). Crime Prediction and Analysis Using Machine Learning. International Research Journal of Engineering and Technology, 5(9), 1037–1042. www.irjet.net

Brahan, J. W., Lam, K. P., Chan, H., & Leung, W. (1998). AICAMS: artificial intelligence crime analysis and management system. Knowledge-Based Systems, 11(5–6), 355–361. https://doi.org/10.1016/S0950-7051(98)00064-1

Bucci, B. A., & Vipperman, J. S. (2007). Performance of artificial neural network-based classifiers to identify military impulse noise. The Journal of the Acoustical Society of America, 122(3), 1602–1610. https://doi.org/10.1121/1.2756969

Chidambaram, B., Pigg, P., Horn, P. L., Talbot, M. A., Gilbertson, D. D., & Cerise, K. C. (2006). Recorders, Reasoners and Artificial Intelligence – Integrated Diagnostics on Military Transport Aircraft. 2006 IEEE Aerospace Conference, 2006, 1–9. https://doi.org/10.1109/AERO.2006.1656074

Cioppa, T. M., Lucas, T. W., & Sanchez, S. M. (2004). Military Applications of Agent-based Simulations. Proceedings of the 36th Conference on Winter Simulation, 171–180. http://dl.acm.org/citation.cfm?id=1161734.1161774

Cirincione, G., Pham, T., Ladas, A. P., Stanton, B., & Fischer, G. (2019). Design and implementation of the U.S. Army Artificial Intelligence Innovation Institute. In T. Pham (Ed.), Artificial Intelligence and Machine Learning for Multi-Domain Operations Applications (Vol. 11006, p. 10). SPIE. https://doi.org/10.1117/12.2524026

Fleming, S. ~W., Titus, M., Watson, J. ~R., & Doring, D. (2019). Technology Demonstration for One-Week-Ahead Forecasting of Toxic Algal Blooms in the US Army Corps of Engineers Reservoir at Detroit Lake using Machine Learning. AGU Fall Meeting Abstracts, 2019, GH44A-12.

Fonseka, T. M., Bhat, V., & Kennedy, S. H. (2019). The utility of artificial intelligence in suicide risk prediction and the management of suicidal behaviors. Australian & New Zealand Journal of Psychiatry, 53(10), 954–964. https://doi.org/10.1177/0004867419864428

Fossaceca, J. M., & Young, S. H. (2018). Artificial intelligence and machine learning for future army applications. In T. Pham, M. A. Kolodny, & D. M. Wiegmann (Eds.), Ground/Air Multisensor Interoperability, Integration, and Networking for Persistent ISR IX (Vol. 1063507, Issue May 2018, p. 6). SPIE. https://doi.org/10.1117/12.2307753

Garay, F., Gutiérrez, D., Martínez, F., Lombana, D., Ibagué, H., & Jiménez, J. (2019). Comparative Analysis Between Training Methods for a Fuzzy Inference System Potentially Applicable to the Assessment of the Health Status of the Spine in Military Personnel. In J. C. Figueroa-García, M. Duarte-González, S. Jaramillo-Isaza, A. D. Orjuela-Cañon, & Y. Díaz-Gutierrez (Eds.), Applied Computer Sciences in Engineering (pp. 451–471). Springer International Publishing. https://doi.org/10.1007/978-3-030-31019-6_39

Hardy, T. (2001). IA: Inteligencia Artificial. Polis, Revista de La Universidad Bolivariana, 1(2), 0. http://redalyc.org/articulo.oa?id=30500219

Jaszlics, I. J., Jaszlics, S. L., & Jones, S. H. (1994). Automated battlefield simulation command and control using artificial neural networks. Fifth Annual Conference on AI, and Planning in High Autonomy Systems, 100–105. https://doi.org/10.1109/AIHAS.1994.390494

Joh, E. E. (2017). Artificial Intelligence and Policing: First Questions. Seattle University Law Review, 41. https://heinonline.org/HOL/Page?handle=hein.journals/sealr41&id=1161&div=&collection=

Kessler, R. C., Stein, M. B., Petukhova, M. V., Bliese, P., Bossarte, R. M., Bromet, E. J., Fullerton, C. S., Gilman, S. E., Ivany, C., Lewandowski-Romps, L., Millikan Bell, A., Naifeh, J. A., Nock, M. K., Reis, B. Y., Rosellini, A. J., Sampson, N. A., Zaslavsky, A. M., Ursano, R. J., & Army STARRS Collaborators. (2017). Predicting suicides after outpatient mental health visits in the Army Study to Assess Risk and Resilience in Servicemembers (Army STARRS). Molecular Psychiatry, 22(4), 544–551. https://doi.org/10.1038/mp.2016.110

Knox, W. B., & Stone, P. (2009). Interactively shaping agents via human reinforcement. Proceedings of the Fifth International Conference on Knowledge Capture - K-CAP ’09, 9. https://doi.org/10.1145/1597735.1597738

Kott, A. (2018). Intelligent Autonomous Agents are Key to Cyber Defense of the Future Army Networks. In The Cyber Defense Review (Vol. 3, Issue 3, pp. 57–70). Army Cyber Institute. https://www.jstor.org/stable/10.2307/26554997

Kumar, M., Athulya, S., Minu M.B., M., Vinodini M.D., V., Lakshmi K.G., A., Anjana, S., & Manojkumar, T. K. (2018). Forecasting of Annual Crime Rate in India: A case Study. 2018 International Conference on Advances in Computing, Communications and Informatics (ICACCI), 2087–2092. https://doi.org/10.1109/ICACCI.2018.8554422

Mcclendon, L., & Meghanathan, N. (2015). Using machine learning algorithms to analyze crime data. Machine Learning and Applications: An International Journal (MLAIJ), 2(1), 1–12. https://doi.org/10.5121/mlaij.2015.2101

Pinsky, J. A., & Lynn, J. T. (1992). Artificial intelligence for data reduction in key management applications. Proceedings - IEEE Military Communications Conference MILCOM, 1992-October, 1111–1115. https://doi.org/10.1109/MILCOM.1992.244113

Rahul Shah, R. (2020). Crime Prediction Using Machine Learning. https://www.kaggle.com/wosaku/crime-in-vancouver

Redmond, M., & Baveja, A. (2002). A data-driven software tool for enabling cooperative information sharing among police departments. European Journal of Operational Research, 141(3), 660–678. https://doi.org/10.1016/S0377-2217(01)00264-8

Schutzer, D. (1983). Applications of Artificial Intelligence Military Communications. IEEE MILITARY COMMUNICATIONS CONFERENCE, 786–790. https://vdocuments.mx/documents/ieee-1983-ieee-military-communications-conference-washington-dc-usa-19831031-1983112-58c3518f8656e.html

Sen, D., Chakrabarti, R., Chatterjee, S., Grewal, D. S., & Manrai, K. (2019). Artificial intelligence and the radiologist: the future in the Armed Forces Medical Services. Journal of the Royal Army Medical Corps, jramc-2018-001055. https://doi.org/10.1136/jramc-2018-001055

Shah, V., & Buckner, G. D. (1988). Potential defense applications of expert systems. IEEE Aerospace and Electronic Systems Magazine, 3(2), 15–21. https://doi.org/10.1109/62.849

Smith, R. (2010). The Long History of Gaming in Military Training. Simulation & Gaming, 41(1), 6–19. https://doi.org/10.1177/1046878109334330

Stanton, B., Cirincione, G., Pham, T., Fischer, G., Ross, S., Whipps, G., Wiegmann, D., & Bennett, K. (2020). US Army Artificial Intelligence Innovation Institute (A2I2) Aiding Multi-Domain Operations (MDO). https://apps.dtic.mil/sti/citations/AD1104260

Stivoric, J. M. K., Boehmke, S. K., Teller, E., & Kasabach, C. (2017). Apparatus for detecting human physiological and contextual information (Patent No. US20170112391A1). https://patents.google.com/patent/US20170112391A1/en

Wilson, B. (2017). Interfacing force on force and communications models MANA and JNE. RAND

Date de modification :