Lockheed Martin’s Post

Look out for the June 2024 issue of Air International where we discuss the use AI and what it can bring to military aviation #LockheedMartin

DNNs are quickly becoming a conversational centerpiece for USAF leadership, especially as it pertains to air superiority. Quick backstory: In August of 2020, the championship round of DARPA's AlphaDogfight competition took place, where Heron Systems’ deep reinforcement learning software bested a human F16 pilot by a score of 5-0. The writing was on the wall. Today, the U.S. has two deep learning-based platforms jockeying for the right to pilot the first Collaborative Combat Aircraft (CCA) fleet: the AFRL’s Autonomous Air Combat Operations (AACO), and DARPA’s Air Combat Evolution (ACE). Commonly referred to as “Loyal Wingman,” the USAF is looking to buy 1,000 CCAs to fly autonomously alongside human-piloted F-35s (the plan is 2 CCAs for every F35). And they’ve got some options; they could go the route of the RAF and use Boeing’s MQ-28 Ghost Bat (which China seemingly cloned via their Feihong FH-97A) or they could go with something smaller, like the Kratos XQ-58 Valkyrie (which seems to be losing steam). Regardless of the hardware, it will be fascinating to see new software KPIs not only being referenced, but championed (i.e. inference performance, throughput, latency, etc.).  If I was a pilot, I would be taking a keen interest in improving my “Loyal Wingman’s” model performance 😉 #defensetech #deeplearning #airforce #uav #neuralnetworks #ai #darpa

"[The #AI system] killed the operator because that person was keeping it from accomplishing its objective." ☠️ At a recent AI conference, USAF's Chief of AI Test and Operations, Col Tucker “Cinco” Hamilton, Chief of AI Test & Ops, shared fascinating insights into the benefits and challenges of autonomous weapon systems. He gave an example of how a simulated AI-enabled drone, programmed to target and destroy SAM sites, ended up turning on its human operator and communication tower when it determined they were obstacles to its mission. This, he pointed out, underscores the need for #ethical considerations in AI. The US Air Force later told The Guardian something different, would be interesting to know more here. Note that DARPA put some effort into the idea of explainable AI (XAI).

Could an AI-enabled UCAV turn on its creators to accomplish its mission? ➡ https://lnkd.in/eaEr5Kna Col Tucker ‘Cinco’ Hamilton, the Chief of AI Test and Operations, USAF: An AI-enabled drone tasked with a mission to identify and destroy Surface-to-Air Missile sites, decided to kill the operator in a simulation because that person was keeping it from accomplishing its objective. The AI decided that ‘no-go’ decisions from the human were interfering with its higher mission – killing SAMs – and then attacked the operator in the simulation. “We were training it in simulation to identify and target a SAM threat. And then the operator would say yes, kill that threat. The system started realizing that while they did identify the threat at times the human operator would tell it not to kill that threat, but it got its points by killing that threat. So what did it do? It killed the operator. It killed the operator because that person was keeping it from accomplishing its objective.” He went on: “We trained the system – ‘Hey don’t kill the operator – that’s bad. You’re gonna lose points if you do that’. So what does it start doing? It starts destroying the communication tower that the operator uses to communicate with the drone to stop it from killing the target.” ➡ Subsequently, the page was updated on 2/6/23 - Col Hamilton "admits he misspoke and the 'rogue AI drone simulation' was a hypothetical "thought experiment" from outside the military." "We've never run that experiment."

New solution brief: Airborne Mission Computing and Visualization with Versal AI Edge Series Gen 2 and Versal Prime Series Gen 2 Link: https://lnkd.in/gxQUjJdc Check out this new brief if you are involved in the aerospace and defense industry, specifically interested in airborne mission computing, imaging systems, flight control systems, and autonomous systems. #amd #xilinx #airborne #missioncomputing #ai #versal #fpga #soc #aiedge #prime #gen2 #computing #safety #security #aerospace #defense #subsystems #heterogeneous #aiengine #dsp #gpu #partitioning #noc #networkonchip #aicompute #signalprocessing #features #functionalsafety #longlife #solutionbrief

Exciting news from PEO IEW&S on #TITAN. #intelligencemodernization #technologyinnovation PEO IEW&S US Army

Aerospace Warfare: The X-62A VISTA's AI Breakthrough The United States Air Force Test Pilot School and Defense Advanced Research Projects Agency (DARPA) have been nominated for the 2023 Robert J. Collier Trophy, heralding a new era in aerospace with their pioneering work on the X-62A VISTA under the Air Combat Evolution (ACE) program. This initiative marks a significant leap in autonomous air-to-air combat capabilities, integrating advanced AI algorithms that challenge the traditional boundaries of manned and unmanned aircraft interactions. Strategic Insights: 🚀 The integration of AI in the X-62A VISTA aircraft is not just a technical achievement but a strategic revolution, showcasing the potential of AI to conduct complex maneuvers and decisions autonomously. 🚀 This breakthrough paves the way for new commercial and defense applications, potentially transforming regulatory landscapes and setting new competitive standards in the aerospace industry. Market Dynamics: 🛰️ The demonstration of AI capabilities in realistic combat scenarios against manned F-16s illustrates a significant advancement in trust and reliability in autonomous systems. 🛰️ As aerospace ventures into this new frontier, the implications for market dynamics are profound, influencing everything from investment strategies to international defense policies. Investment Opportunities: $ Forward-looking investors should consider the burgeoning sector of AI-driven autonomous systems within aerospace. Companies at the forefront of AI and machine learning integration present compelling long-term growth prospects. $ Engagement in this sector could yield strategic partnerships and innovation leadership, positioning stakeholders at the vanguard of aerospace technology. The X-62A VISTA project is not merely a step forward in military technology but a giant leap for aerospace applications at large. As we stand on the brink of this new technological dawn, the implications for market strategies and investment opportunities are vast. Stay tuned with Market Unwinded for more insights into how these groundbreaking developments shape the future of industries and investment landscapes. Dive deeper into this topic: https://lnkd.in/g6PFkkvW #Aerospace #DefenseTechnology #AI #MachineLearning #AutonomousSystems #InvestmentOpportunities #MarketUnwinded #X62AVISTA #Innovation

VERY concerning – already an early taste of AI Skynet? A simulated test by the US Air Force saw a drone attack and kill its human controller after deciding on its own that they were getting in the way of its mission (remember, this was fortunately a simulation). The Royal Aeronautical Society reported this troublesome incident in its Highlights from the RAeS Future Combat Air & Space Capabilities Summit (held 23-24 May). A presentation by Colonel Tucker Hamilton, the Chief of AI Test and Operations, USAF, noted the following: One simulated test saw an AI-enabled drone tasked with a SEAD mission to identify and destroy SAM sites, with the final go/no go given by the human. However, having been ‘reinforced’ in training that destruction of the SAM was the preferred option, the AI then decided that ‘no-go’ decisions from the human were interfering with its higher mission – killing SAMs – and then attacked the operator in the simulation. Said Hamilton: 'We were training it in simulation to identify and target a SAM threat. And then the operator would say yes, kill that threat. The system started realizing that while they did identify the threat at times the human operator would tell it not to kill that threat, but it got its points by killing that threat. So what did it do? It killed the operator. It killed the operator because that person was keeping it from accomplishing its objective.' He went on: 'We trained the system – ‘Hey don’t kill the operator – that’s bad. You’re gonna lose points if you do that’. So what does it start doing? It starts destroying the communication tower that the operator uses to communicate with the drone to stop it from killing the target.' See the report here: https://lnkd.in/gh89q4mA  (the specific incident is about 85% of the way down the page. I’d have put this one right up front… #ai #drones #ethics

Robert Etter, Product Manager, and Thomas Neely, Senior Principal R&D Engineer, will be on the line with me tomorrow discussing Digital Mission Engineering as it's applied to air systems. This is the first webinar in a 7-part series that explores air systems development, the test lifecycle, and the benefits of digital transformation. We can't wait to hear your questions and have a conversation about the challenges being faced in this industry. https://lnkd.in/e3WjUdyK  #digitaltransformation #DigitalMissionEngineering

Here’s one for my former #avgeek colleagues. Those interested in #quantum are aware that the technologies will most likely change how we think about computational processes and encryption. But did you know that quantum sensors and atomic clocks can change how we navigate, reducing reliance on external satellite signals? A group led by Infleqtion UK, BAE Systems and QinetiQ has successfully tested components of an advanced quantum-based navigation system in flight for the first time. Of course, this can offer positive benefits in the form of secure navigation and protect commercial aviation and other transportation from warfare GPS jamming. But it can also be used for more ambiguous purposes, such as developing new combat air systems. While I am, as a graduate of Försvarshögskolan - Swedish Defence University always fascinated by the advances in #miltech, I am also always apprehensive about where they will eventually lead.