(C) Daily Kos
This story was originally published by Daily Kos and is unaltered.
 . . . . . . . . . .



The Impact of AI Transformers on Military Planning-Ukrainian Iterations [1]

['This Content Is Not Subject To Review Daily Kos Staff Prior To Publication.', 'Backgroundurl Avatar_Large', 'Nickname', 'Joined', 'Created_At', 'Story Count', 'N_Stories', 'Comment Count', 'N_Comments', 'Popular Tags']

Date: 2023-04-19  

Self Organizing Ukrainian Sunflowers

One can’t understand the impact of artificial intelligence on military planning and operations in Ukraine without first understanding the role of AI transformers. According to Wikipedia:

“A transformer is a deep learning model that adopts the mechanism of self-attention, differentially weighting the significance of each part of the input (which includes the recursive output) data. It is used primarily in the fields of natural language processing (NLP)[1] and computer vision (CV).[2]”

Transformer models take what were formerly separate and distinct aspects of information processing such as time, sequence, space, geographic/terrain, vision, logic, and language analysis, image acquisition and object recognition, robotics, network and systems description, simulations and analysis, command formulation, control and iteration and evaluates their relative weights and importance in accomplishing objectives.

One thought leader in applying transformer models in military simulations is Roger Smith. He has recently written about how our military services have invested heavily in creating, deploying, and training with large simulation systems. For the non-expert, the scope of the systems that he describes is breathtaking. However he explains the process that is required to create and field these simulators as very one-way and linear and begins with organizations like the Army Training and Doctrine Command (TRADOC), proceeds to acquisition by the Program Executive Office for Simulation, Training and Instrumentation (PEO-STRI), is deployed to Combat Training Centers (CTC), with the post-exercise analysis delivered to the Center for Army Lessons Learned (CALL). He notes that this linearity may align poorly to what is becoming possible given the quantum changes he foresees from the next generation of AI transformer and deep learning technologies. He highlights that the DOD acquisition process, which creates new simulators just as it does helicopters, ground vehicles, and communication networks, is complex and difficult to explain, but needs to rapidly adapt in order to take full advantage of the new tools.

Roger D. Smith. 2022. Applying AI Deep Learning to the DOD’s big simulation and training projects. In SIGSIM Conference on Principles of Advanced Discrete Simulation (SIGSIM-PADS ’22), June 08–10, 2022, Atlanta, GA, USA. ACM, New York, NY, USA, 5 pages. https://doi.org/10.1145/3518997.3534118 link at https://web.archive.org/web/20220617011907id_/https://dl.acm.org/doi/pdf/10.1145/3518997.3534118

Like any good computer scientist and systems analyst Mr Smith explains to the end users and procurement officials things like the following:

“military training and simulation process described . . . contains two phases that are extremely data intensive, making them potential candidates to benefit from deep learning algorithms.” “The military training enterprise . . . is just beginning to realize [big data’s] potential. For example, real time data processing with deep learning could be used to track classify the activities of complex groups of entities, define geographic areas according to military control, track threat areas of influence, and translate specific orders and reports into larger objectives.” “The Prepare phase includes the creation of very large scenario databases that describe in minute detail the characteristics and capabilities of all types of military equipment, organizations, and command structures. Each scenario is a geographic instantiation of the forces that will be involved in the training event. Hundreds or thousands of entities or aggregate units must be placed on a battlefield, arranged according to current military doctrine, and applied to unique terrain features. This is an extremely tedious and laborious process. It is so laborious that, in many cases, scenarios are reused from previous events with minimal modifications. [and deep learning can change this]” “Deep learning algorithms offer the potential to generate unique laydowns within the constraints of doctrine and terrain. These algorithms can be expected to discover patterns that fit these constraints in unique ways that have not be[en] considered or used by humans in the past.” “The Execute phase of the training process is already rich with mathematic and symbolic AI algorithms that support the process of running an exercise and stimulating the training audience. Deep learning has a place among these as well. These methods can be used to analyze the patterns of enemy unit locations and movement, and to select a response that is optimal for countering the opponent. The process is similar to playing popular games like Chess and Go, while discovering powerful moves that were not previously explored by human players.”

Given the necessarily secretive nature of military planning and operations, most people can only conjecture about how the illegal Russian invasion of Ukraine has accelerated the adoption of big data and artificial intelligence in planning for and executing Ukraine’s defense and hoped for victory, a la Smith’s 2022 suggestions. However, one example (reported by the very excellent “Reporting From Ukraine” YouTube channel) from the very effective Ukrainian counterattack on Russian military forces yesterday at Bakhmut may be an indication of how rapidly things have changed.

https://youtube.com/watch?v=M8C2yyPrzag&feature=share

[END]
---
[1] Url: https://www.dailykos.com/stories/2023/4/19/2164719/-The-Impact-of-AI-Transformers-on-Military-Planning-Ukrainian-Iterations

Published and (C) by Daily Kos
Content appears here under this condition or license: Site content may be used for any purpose without permission unless otherwise specified.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/dailykos/