Assessment of Civilian Next-Generation Knowledge Management Systems for Managing Civil Information

Assessment of Civilian Next-Generation Knowledge Management Systems for Managing Civil Information

Ray K. Ragan

This article is published as part of the Small Wars Journal and Divergent Options Writing Contest which ran from March 1, 2019 to May 31, 2019.  More information about the writing contest can be found here.

Current Civil Information Management Systems are not taking advantage of the leaps of technology in knowledge management, specifically in the realm of predictive analytics, Natural Language Processing, and Machine Learning. This creates a time cost that commanders must pay in real-time in their operating environment, particularly felt in small wars. This cost also diverts resources away from direct mission-enabling operations.

Text:  Currently Civil Information Management (CIM) systems employed by the U.S. Military are not keeping pace with the current revolution seen in civilian next-generation knowledge management systems (KMS)[1][2]. These KMS are possible through the convergence of modern computing, predictive analytics, Natural Language Processing (NLP), and Machine Learning (ML)[3]. This CIM limitation is unnecessary and self-imposed as a KMS offers persistent and progressing inputs to the common operating picture. This assessment explores how civilian business harnessed this revolution and how to apply it to CIM.

Generally, CIM represents the operational variables (OV) of an operational environment (OE) and as practiced today, resides in the domain of information rather than knowledge[4]. The DIKW pyramid framework, named for its Data, Information, Knowledge, Wisdom structure informs the structure of learning[5]. Further, one can infer that traversing each step represents time and effort, a price paid by commanders in real-time during operations. Small wars demand speed and agility. Current CIM takes time to gather data, input it into a database, run queries, overlay on maps, and eventually infer some knowledge to inform decision making by the commander[6]. 

Using the 1999-invented Cynefin Framework to aid decision-making, commanders needlessly leave many of the OVs in the chaotic domain[7]. Moving from the chaotic to the complex domain the OVs must come from a KMS that is persistent and automatically progressing. Current CIMs do not automatically update by gathering information from public sources such as broadcast, print, and digital that are digitized with NLP and speech/text analytics[8].   Instead, human operators typically located in the OE, manually update these sources. Because of this, today’s CIMs go stale after the operators complete their mission or shift priorities, making what information was gathered devolve to historic data and the OE fog of war revert to chaos[9].

The single biggest advantage a quality KMS provides to a commander is time and decision-making in the OE[10]. Implemented as a simple search engine that is persistent and progressing for all OEs, would mean a commander does not need to spend operational time and effort on basic data gathering missions. Rather, a commander can focus spending operational resources on direct mission-enabling operations. Enticingly, this simple search engine KMS allows for the next advancement, one that businesses around the world are busily employing – operationalizing big data.

Business systems, such as search engines and other applications scour open sources like in court records and organizes them through a myriad of solutions. Data organized through taxonomy and algorithms allow businesses to offer their customers usable information[11]. The advent of ML permits the conversion of information to knowledge[12]. Civilian businesses use all these tools with their call centers to not only capture what customers are saying, but also the broader meta conversation: what most customers are not saying, but revealing through their behavior[13]. 

This leap in application of informatics, which civilian business use today, is absent in today’s CIM systems. The current model of CIM is not well adapted for tomorrow’s battlefield, which will almost certainly be a data-rich environment fed by robotics, signals, and public information[14]. Even the largest team of humans cannot keep up with the overwhelming deluge of data, let alone conduct analysis and make recommendations to the commander of how the civilian terrain will affect his OE[15].

In civilian business, empiricism is replacing the older model of eminence-based decision-making. No longer is it acceptable to take the word of the highest-paid person’s opinion, business decisions need to have evidence, especially at the multi-billion dollar level company level[16]. KMS enables for hypothesis, experimentation, and evidence. Applied in the civilian terrain, if the hypothesis is that by drilling a well reduces insurgency, a global KMS will reveal the truth through the metrics, which cannot be influenced, as former-U.S. Secretary of State Condoleezza Rice criticized[17]. 

Using text preprocessing with speech analytics and NLP, the KMS would solve an OE problem of data quality, as operators when supplementing the KMS with OE reports, would use speech whenever possible. This overcomes a persistent problem of garbage in and garbage out that plagues military and business systems alike. Rather than re-typing the field notes into a form, the human operator would simply use an interactive spoken dialog for input where feasible[18].

A persistent and progressive KMS also addresses a problem with expertise. During Operation Iraqi Freedom, the U.S. State Department could not find enough experts and professionals to fill the voids in transitional governance. This problem was such that then-Secretary of Defense Robert Gates volunteered to send Department of Defense civilians in their place[19]. With a KMS, commanders and policymakers can draw on a home-based cadre of experts to assess the data models of the KMS and offer contextualized insights into the system to commanders in the field.

As the breadth and quality of the data grows, system administrators can experiment with new algorithms and models on the data in a relentless drive to shorten OV-derived insights into operations planning. Within two years, this KMS data would be among the richest political science datasets ever compiled, inviting academia to write new hypothetical models and experiment. In turn, this will assist policy makers in sensing where new sources of instability emerge before they reveal themselves in actions[20].

“How do you put the genie of knowledge back in the bottle,” P. W. Singer rhetorically asked[21] in his book, Wired for War about the prospect of a robotic, data-enabled OE. This genie will not conveniently return to his bottle for robotics or data, instead commanders and policy makers will look to how to manage the data-enabled battlefield. While it may seem a herculean task to virtually recreate OEs in a future KMS, it is a necessary one. Working through the fog of war with a candle and ceding OVs to chaos is no longer acceptable. Civilian business already addressed this problem with next-generation knowledge management systems, which are ready for today’s OE.

End Notes

[1] APAN Staff (n.d.) Tools. Retrieved May 9, 2019, from https://www.apan.org/(S(12adofim0n1ranvobqiyfizu))/pages/tools-communities

[2] Williams, Gregory (2016, December 2). WFX 16 tests Civil Affairs Soldiers. Retrieved May 12, 2019, from https://www.dvidshub.net/news/189856/wfx-16-tests-civil-affairs-soldiers

[3] Szilagyi and P. Wira (2018) An intelligent system for smart buildings using machine learning and semantic technologies: A hybrid data-knowledge approach, 2018 IEEE Industrial Cyber-Physical Systems (ICPS), St. Petersburg, pp. 22-24.

[4] Chief, Civil Affairs Branch et al. (2011). Joint Civil Information Management Tactical Handbook, Tampa, FL, pp. 1-3 – 2-11.

[5] Fricke, Martin (2018, June 7). Encyclopedia of Knowledge Organization: Knowledge pyramid The DIKW hierarchy. Retrieved May 19, 2019, from http://www.isko.org/cyclo/dikw

[6] Chief, Civil Affairs Branch et al. (2011). Joint Civil Information Management Tactical Handbook, Tampa, FL, pp. 5-5, 5-11.

[7] Kopsch, Thomas and Fox, Amos (2016, August 22). Embracing Complexity: Adjusting Processes to Meet the Challenges of the Contemporary Operating Environment. Retrieved May 19, 2019, from https://www.armyupress.army.mil/Journals/Military-Review/Online-Exclusive/2016-Online-Exclusive-Articles/Embracing-Complexity-Adjusting-Processes/

[8] APAN Staff (n.d.) Tools. Retrieved May 9, 2019, from https://www.apan.org/(S(12adofim0n1ranvobqiyfizu))/pages/tools-communities

[9] Neubarth, Michael (2013, June 28). Dirty Email Data Takes Its Toll. Retrieved May 20, 2019, from https://www.towerdata.com/blog/bid/116629/Dirty-Email-Data-Takes-Its-Toll

[10] Marczewski, Andrzey (2013, August 5). The Effect of Time on Decision Making. Retrieved May 20, 2019, from https://www.gamified.uk/2013/08/05/the-effect-of-time-on-decision-making/

[11] Murthy, Praveen et al. (2014, September). Big Data Taxonomy, Big Data Working Group, Cloud Security Alliance, pp. 9-29.

[12] Edwards, Gavin (2018, November 18). Machine Learning | An Introduction. Retrieved May 25, 2019, from https://towardsdatascience.com/machine-learning-an-introduction-23b84d51e6d0

[13] Gallino, Jeff (2019, May 14). Transforming the Call Center into a Competitive Advantage. Retrieved May 25, 2019, from https://www.martechadvisor.com/articles/customer-experience-2/transforming-the-call-center-into-a-competitive-advantage/

[14] Vergun, David (2018, August 21). Artificial intelligence likely to help shape future battlefield, says Army vice chief.  Retrieved May 25, 2019, from https://www.army.mil/article/210134/artificial_intelligence_likely_to_help_shape_future_battlefield_says_army_vice_chief

[15] Snibbe, Alana Conner (2006, Fall). Drowning in Data. Retrieved May 25, 2019, from https://ssir.org/articles/entry/drowning_in_data

[16] Frizzo-Barker, Julie et al. An empirical study of the rise of big data in business scholarship, International Journal of Information Management, Burnaby, Canada, pp. 403-413.

[17] Rice, Condoleezza (2011) No Higher Honor. New York, NY, Random House Publishing, pp. 506-515.

[18] Ganesan, Kavita (n.d.) All you need to know about text preprocessing for NLP and Machine Learning. Retrieved May 25, 2019, from https://www.kdnuggets.com/2019/04/text-preprocessing-nlp-machine-learning.html

[19] Gates, Robert (2014). Duty. New York, NY, Penguin Random House Publishing, pp. 347-348.

[20] Lasseter, Tom (2019, April 26). ‘Black sheep’: The mastermind of Sri Lanka’s Easter Sunday bombs. Retrieved May 25, 2019, from https://www.reuters.com/article/us-sri-lanka-blasts-mastermind-insight/black-sheep-the-mastermind-of-sri-lankas-easter-sunday-bombs-idUSKCN1S21S8

[21] Singer, Peter Warren (2009). Wired for War. The Penguin Press, New York, NY, pp. 11.