NDIA’s Virtual SOFIC 2021 comes to a close and what an event it was in the virtual domain. NDIA, USSOCOM and the many corporate sponsors deserve to be congratulated for what can only be described as a brilliantly planned, choreographed and executed conference. The amount of information presented was overwhelming but it communicated a direction that can only be described as Star Wars. Yes, the usual mobility, air lift, SOF Operator requirements were discussed, but at the heart of SOF of the future is artificial intelligence, machine learning, neural networks, quantum computing and advanced sensors. SOF focus is no longer on simply shortening the kill chain, they are zeroed-in on confronting and defeating peer competitors, and that will require resources and technology the likes of which I can only imagine. Were I a computer science, engineering, physics and biotechnology major USSOCOM would be the place I want to be a part of and I encourage students in those disciples to consider the organization as your career choice.

It’s important to communicate the 30,000 foot view of what USSOCOM’s mid to longer term objects are and to do that I’ve created what I’ll call an executive cartoon for you. I’ll do my best to walk you through the nuances.

AI, Deep Learning, and Neural Networks Explained

A neural network is a strategy that emulates the human nervous system. It eliminates single points of failure and is self healing. The engine behind the network is artificial intelligence, a series of algorithms that can operate independently or interactively. For the sake of discussion lets assume that what I show as inputs are sensors and that each of those sensors could be SOF operator worn, UAV, UGV and even UUV assets, satellites, submarines, or any other device such as a tagging and tracking sensor. They’re deployed and collect information. That information can be shared through the network where AI software can process it and route it to the appropriate destination. Now imagine that through machine learning that sensor can be taught to recognize a threat or event. Further imagine that the sensor has its own AI capability. Each sensor can then direct a specific amount of data to the ultimate source. Here’s an example.

An operator worn sensor detects the presence of an r.f. emission that it recognizes as a threat and without operator intervention directs a UAV to conduct a strike on the source of that r.f. emission.

SOCOM’s objective is precisely that, to deploy advanced sensors that can be reconfigured over the network to meet mission objectives and communicate through software defined waveforms, that are difficult to detect or jam, on a neural network that may also include space assets. The DOD has already deployed neural networks but SOF requirements will be significantly higher. If successful, what SOF is doing could readily be transferred to law enforcement and used in commercial applications. This is an ambitious plan that will require human capital, out of the box thinking and a tremendous amount of technology. I can’t wait to see the finished product. –SP

This entry was posted in SOFIC 2021, SOFIC 2021 (virtual) and tagged , , , , . Bookmark the permalink.