www.c4isrnet.com Open in urlscan Pro
2a02:26f0:2780:6::214:f58c  Public Scan

Form analysis
0 forms found in the DOM

Text Content

Sections
AI & MLUnmannedBattlefield TechSpaceElectronic WarfareCyberIndustryOpinion

 * AI & ML
 * Unmanned
    * Robotics

 * Battlefield Tech
    * Hypersonics
    * JADC2 & Comms
    * IT/Networks
    * Space

 * Information Warfare
 * Electronic Warfare
 * Cyber
 * Industry
    * 3D Printing

 * Videos
 * Opinion
    * The Compass

 * Native
 * Events Calendar
 * NewslettersOpens in new window
 * 

Featured:

Coverage: AUSA Annual
Whitepaper: Guarding Against AI Threats
Whitepaper: Cybersecurity in the Quantum Era




AI


INFORMATION SUPERIORITY AND THE RACE FOR GLOBAL AI LEADERSHIP

By Col. Doug Drakeley (Ret.)
 Dec 6, 2022

Thanks to advances in AI, human-agent teaming and machine learning, soldiers
will provide commanders with real-time information about the enemy gathered from
a variety of different sources, including possible courses of action, which will
help them to make better decisions in battle. (Army)

The first offset strategy centered around nuclear capabilities, the second on
stealth technology. Now, with competing powers once again approaching parity
with the U.S. in critical military capabilities, the Department of Defense is
pursuing a third offset strategy — one that puts Artificial Intelligence at the
forefront of U.S. defense policy.

This new offset strategy aims to improve the U.S.’s information management
capabilities through the adoption of cutting-edge technologies, including AI
systems. Because in a data-driven military landscape, the strategic advantage
lies with entities that can collect, process, and respond to information faster
and more accurately than their adversaries.



Speed and accuracy are often trade-offs with AI systems — particularly for
complex, multi-faceted enterprises. DoD efforts to assimilate Joint Force
networks for greater operational speed and reliability are underway. However,
legacy AI technologies lack the computing power to process information
effectively across domains and data sources, especially at scale.

So, how can the U.S. advance its AI capabilities and maintain a strategic edge?
Recent developments in AI point to multimodal capability as a key
differentiator.


PAST LIMITATIONS, FUTURE POTENTIAL

Although there’s debate about whether the U.S. still boasts better AI
capabilities than its peers and near-peers, it’s undeniable that competitors are
closing the gap. And even if the DoD still holds the upper hand in AI, its
advantage is no longer enough to rely on technological superiority as a
peace-keeping measure.



The problem isn’t that the DoD lacks the personnel or computing power to outpace
other nations. The problem is that U.S. AI networks struggle to convert massive
amounts of data into usable insights with enough speed and accuracy to project
superiority.

This challenge is difficult to overcome with conventional AI and machine
learning systems, which lack the general-purpose capabilities to integrate
unlike data types and seamlessly manage information across domains. The
objectives of the U.S.’s third offset require the elimination of these barriers
in support of more efficient data processing and, subsequently, more informed
decision-making on the part of military personnel.

Up to now, the DOD has undertaken efforts to improve information management
capabilities through the Joint All-Domain Command and Control network, which
addresses long-standing issues from data silos and stovepipe systems. However, a
consolidated Joint Force network can only operate as fast and accurately as AI
systems allow it to — hence the pursuit of superior AI capabilities.


A MODEL FOR LONG-TERM AI LEADERSHIP

The emergence of multimodal AI (also known as foundation models) represents a
significant breakthrough in AI technology for both the private sector and the
military.



While past generations of AI systems relied on task-centric infrastructure —
where each use case required its own model and associated training — multimodal
AI eliminates those rigidities through in-context learning. This learning
structure gives multimodal AI the flexibility to process various data types with
a combination of algorithms, accelerating information collection and processing
across networks for more sophisticated data analysis and decision-making.

Put simply, this multimodal structure generates relevant insights from multiple
data sources much faster — and on a much larger scale — than previously
possible.


MULTIMODAL AI AS THE FIRST LINE OF DEFENSE

The DoD’s ability to use AI to gain full situational awareness through a
multi-domain defense strategy becomes much more robust with the versatility of
multimodal AI. It’s more accurate than conventional models and capable of
zero-shot and few-shot learning. For example, a Contrastive Language-Image
Pre-Training model can classify images from a given set of language-expressed
categories without needing fine-tuning.

The adaptability of multimodal models allows them to cut through the complexity
of the data that’s generated and integrated across domain networks to help
operators understand all available options and inform the best course of action.
If an adversary launches an attack by sea, AI can rapidly determine if the
proper response is to fire missiles, launch fighters, or execute a cyberattack.



Additionally, the development of greater AI functionality will feature an
iterative process. Wide-ranging applications for multimodal AI promise to
enhance human-machine collaboration across all fronts to support more vital
mission capabilities — for personnel on the front lines and in the data centers.

It’s important to note, however, that multimodel systems need to have a large
amount of processing power, sizeable on-chip memory, and enough attached memory
to handle data efficiently. So, an integrated hardware-software systems approach
is necessary to create the right balance of computing, memory, and communication
for data-intensive dataflow operations. Ideally, these systems should be
flexible to handle the inference and incremental training for superior model
creation.


TAKING OUTDATED AI TO TASK

Consider the value of multimodal for intelligence, surveillance, and
reconnaissance data collection in a Joint Force network. Satellite systems
generate immense amounts of audio and visual data for ISR. While task-centric AI
models struggle to interpret unlike data inputs or recognize meaningful patterns
across various data sources, a foundation model functions as an overarching data
processing hub. Within this hub, scalability and contextual learning
capabilities mean multimodal AI can operate with the same computing productivity
as hundreds of task-centric models.

In the case of ISR data, AI systems under a multimodal system can recognize
patterns from audio and visual inputs to identify and flag if, for example,
satellite video footage of an adversary’s tank movements matches radio
frequencies, indicating a mass military mobilization. The AI will quickly and
correctly make this connection and provide its operators with the relevant
insight they need to craft the best response.



The U.S. has used rising military parity as an impetus to seek superior
technological advantages since the Cold War. And in the era of the third offset
strategy, where AI capabilities represent the latest proving ground for
conventional military deterrence, multimodal AI is the cutting-edge innovation
at the center of it all.

Col. Doug Drakeley (Ret.) is an advisory board member and industrial specialist
at SambaNova Systems, a supplier of AI platforms and services based in Palo
Alto, California.

SHARE:

MORE IN AI

TAIWAN TAPS SATELLITE HOOKUPS TO HELP DOWN INVADING DRONES

"THIS WILL HELP PREVENT SYSTEMIC COLLAPSE OF THE CORE COMMUNICATION NETWORK IN
POTENTIAL FUTURE CONFLICTS,” SAID THE CEO OF TRON FUTURE.



HOW AN EXPERIMENT IN NEW JERSEY COULD SHAPE THE ARMY’S FUTURE NETWORK

THROUGH A SERIES OF EXPERIMENTS KNOWN AS NETMODX, THE ARMY SOUGHT TO WRING OUT
INTEGRATED NETWORK CAPABILITIES IN REAL-WORLD CONDITIONS.



FOR REPLICATOR 2, ARMY WANTS AI-ENABLED COUNTER-DRONE TECH

THE ARMY MAY REQUEST APPROVAL FROM CONGRESS TO REPROGRAM FISCAL 2025 FUNDING FOR
INITIAL REPLICATOR 2 SYSTEMS.



LEONARDO DEBUTS DRONE-MOUNTED JAMMER TO BAMBOOZLE AIR THREATS

THE BRITESTORM, WHICH THE FIRM IS SHOWING OFF AT THE AUSA CONVENTION IN
WASHINGTON, WILL WEIGH ABOUT 2.5KG AND OCCUPY THE SPACE OF SIX COKE CANS.



LEONARDO, BLUEHALO DEMO COUNTER-DRONE SYSTEM ON ARMY STRYKER

THE STRYKER-BASED DIRECTED ENERGY COUNTER-UAS PROTOTYPE CAN DETECT, IDENTIFY AND
TRACK MULTIPLE DRONE THREATS AND USE BOTH KINETIC AND NON-KINETIC EFFECTS.



FEATURED VIDEO

What to look for in a home remodeling contractor — Money Minute
Jeanette Mack from Navy Federal Credit Union explains what you need to look for
before you hire a contractor for your next home renovation.

0 of 1 minute, 1 secondVolume 0%

Press shift question mark to access a list of keyboard shortcuts
Keyboard ShortcutsEnabledDisabled
Shortcuts Open/Close/ or ?
Play/PauseSPACE
Increase Volume↑
Decrease Volume↓
Seek Forward→
Seek Backward←
Captions On/Offc
Fullscreen/Exit Fullscreenf
Mute/Unmutem
Decrease Caption Size-
Increase Caption Size+ or =
Seek %0-9

Settings
OffEnglish
Font Color
White

Font Opacity
100%

Font Size
100%

Font Family
Arial

Character Edge
None

Edge Color
Black

Background Color
Black

Background Opacity
50%

Window Color
Black

Window Opacity
0%

Reset
WhiteBlackRedGreenBlueYellowMagentaCyan
100%75%50%25%
200%175%150%125%100%75%50%
ArialCourierGeorgiaImpactLucida ConsoleTahomaTimes New RomanTrebuchet MSVerdana
NoneRaisedDepressedUniformDrop Shadow
WhiteBlackRedGreenBlueYellowMagentaCyan
WhiteBlackRedGreenBlueYellowMagentaCyan
100%75%50%25%0%
WhiteBlackRedGreenBlueYellowMagentaCyan
100%75%50%25%0%
Auto1080p720p540p360p270p180p
Live
00:04
00:57
01:01






 

WHAT TO LOOK FOR IN A HOME REMODELING CONTRACTOR — MONEY MINUTE



MARINES TO GET BIGGEST MARKSMANSHIP OVERHAUL IN 100 YEARS



NEW MARINE RIFLE QUALIFICATIONS HEAD TO THE INFANTRY | DEFENSE NEWS WEEKLY FULL
EPISODE 10.12.24



AN INSIDE LOOK AT AUSA 2024




TRENDING NOW

 1. HOW AN EXPERIMENT IN NEW JERSEY COULD SHAPE THE ARMY’S FUTURE NETWORK
    
    

 2. FOR REPLICATOR 2, ARMY WANTS AI-ENABLED COUNTER-DRONE TECH
    
    

 3. AIR FORCE’S INFORMATION WARFARE HUB SEEKS HIGH-TECH SENSORS, AI TOOLS
    
    

 4. WHY SPACE FORCE CHOSE COMMERCIAL FIRMS TO BUILD ITS NEW GROUND SYSTEM
    
    

 5. PALANTIR WINS CONTRACT TO EXPAND ACCESS TO PROJECT MAVEN AI TOOLS
    
    



Facebook pageTwitter feedRSS feed

C4ISRNet © 2024


TERMS OF USE

 * Privacy Policy
 * Terms of Service


GET US

 * NewslettersOpens in new window
 * EventsOpens in new window
 * RSS FeedsOpens in new window


CONTACT US

 * Advertise
 * General Contacts, Subscription Services
 * Editorial Staff


ABOUT US

 * About Us
 * CareersOpens in new window


search by queryly Advanced Search