Ukrainian brigade says 19 killed in Russian missile strike last week | Reuters

Sur le même sujet

• 

  19 septembre 2018 | International, C4ISR

  Differentiating a port from a shipyard is a new kind of problem for AI

  By: Daniel Cebul It's well known that satellites and other intelligence, surveillance and reconnaissance platforms collect more data than is possible for humans to analyze. To tackle this problem, the Intelligence Advanced Research Projects Activity, or IARPA, conducted the Functional Map of the World (fMoW) TopCoder challenge from July 2017 through February 2018, inviting researchers in industry and academia to develop deep learning algorithms capable of scanning and identifying different classes of objects in satellite imagery. IARPA curated a dataset of 1 million annotated, high-resolution satellite images aggregated using automated algorithms and crowd sourced images for competitors to train their algorithms to classify objects into 63 classes, such as airports, schools, oil wells, shipyards, or ports. Researchers powered their deep learning algorithms by combining large neural networks, known as convolutional neural networks (CNNs), and computers with large amounts of processing power. The result was a network that, when fed massive amounts of training data, can learn to identify and classify various objects from satellite imagery. By combining a number of these networks into what is called an ensemble, the algorithm can judge the results from each CNN to produce a final, improved result that is more robust than any single CNN. This is how a team from Lockheed Martin, led by Mark Pritt, designed their deep learning algorithm for the challenge. Pritt explained to C4ISRNET, that he and his team developed their CNN using machine learning software and framework from online open source software libraries, such as Tensor Flow. Earning a top five finish, the algorithm designed by Pritt's team achieved a total accuracy of 83 percent, and was able to classify 100 objects per second. Pritt said that with fully functioning algorithm, this software could take an image recognition task that takes a human an hour to complete and reduce the process to a few seconds. The team's algorithm excelled at identifying classes with distinctive features, and successfully matched nuclear power plants, tunnel openings, runways, tool booths, and wind farms with accuracies greater than 95 percent, but struggled with more indiscreet classes such as shipyards and ports, hospitals, office buildings, and police stations. “Usually when you develop an algorithm its nice to see where it succeeds, but you actually learn the most where you look at where the algorithm fails or it doesn't do well,” Pritt said. In trying to decipher why the algorithms struggled, Pritt said the competitors suggested that some objects simply don't have any distinguishing features from the point of view of a satellite image for the algorithms to recognize. “Maybe the most important ingredient you need for these new types of algorithm to work is the dataset because these algorithms require a great amount of data to train on,” Pritt explained. “It's kind of analogous to the way a human will learn in childhood how to recognize things. You need lots of examples of what those things are and then you can start to generalize and make your own judgments,” he said. But even with large amounts of training data that is correctly labeled, it is also possible the deep learning technology of today cannot reach the higher levels of intelligence to recognize nuanced differences. For example, Lockheed Martin's algorithm confused shipyards and ports 56 percent of the time. Pritt said that people “look at an image and they can tell that it's a port or a shipyard, they are usually looking at very subtle things such as if there is a ship in dry dock or if there is a certain type of crane present. They are looking for details in the image that are maybe higher level or more complicated than what these deep learning algorithms can do right now.” However, the fact that these algorithms cannot do everything should not dismiss the significant contribution they could provide to the defense and intelligence community. Hakjae Kim, IARPA's program manager for the fMoW challenge, said the benefits of this technology could extend far beyond faster image processing. “I want to look at it more in the perspective that we can do things we weren't able to do before,” Kim said. “Because its technology that we are now able to do x, y and z, there are more applications you can create because with the human power it is just impossible to do before.” Kim and Pritt stressed managing expectations for CNN-based artificial intelligence. “This is a real technology that will work, but it also has limitations. I don't want to express this technology as a magic box that will just solve everything magically,” Kim said. “I don't want the users in the field to get disappointed by the initial delivery of this technology and say 'Oh, this is another technology that was oversold and this is not something we can use," he added. Part of managing our expectations for AI requires recognizing that although intelligence is in the name, this technology does not think and reason like humans. “A lot of the time we think that because we use the term AI, we tend to think these algorithms are like us, they are intelligent like us,” Pritt said. “And in someways they seem to mimic our intelligence, but when they fail we realize ‘Oh, this algorithm doesn't really know anything, [it] doesn't have any common sense.'” So how are IARPA and Lockheed Martin working to improve their algorithms? For IARPA, Kim's team is working on updating and maintaining their dataset to ensure algorithms have the most up to date information to train on, ultimately making the CNN-based algorithms easier to trust. “[S]ubtle changes in the area mess up the brains of the system and that system will give you a totally wrong answer,” Kim explained. “So we have planned to continuously look over the area and make sure the algorithm we are developing and reassessing for the government to test on and use to be robust enough for their application," he furthered. Work is also underway at American universities. Kim described how a team of researchers at Boston University are using the fMoW dataset and tested algorithms to create heat maps that visualize what part of the image algorithms are using to classify objects. They've found that sometimes it is not the object itself, but clues surrounding the object that aid most in classification. For example a “windmill that actually shows a shadow gives a really good indicator of what that object is,” Kim said. “Shadows show a better view of the object. A shadow is casting the side view of the object over on the ground, so [BU's heat map algorithm] actually points out the shadow is really important and the key feature to make the object identified as a windmill.” But don't expect these algorithms to take away the jobs of analysts any time soon. “I think you still need a human doing the important judgments and kind of higher level thinking,” Pritt said. “I don't think AI will take away our jobs and replace humans, but I think what we have to do is figure out how to use them as a tool and how to use them efficiently, and that of course requires understanding what they do well and what they do poorly," he concluded. https://www.c4isrnet.com/intel-geoint/2018/09/18/differentiating-a-port-from-a-shipyard-is-a-new-kind-of-problem-for-ai

• 

  20 août 2020 | International, Aérospatial, Naval, Terrestre, C4ISR

  The Army is working to see across thousands of miles

  Mark Pomerleau WASHINGTON — The Army's Intelligence, Surveillance and Reconnaissance Task Force is helping the service modernize its ability to see across huge ranges through a layered approach that includes ground, air and space. As geographic boundaries will be blurred in future conflict with sophisticated adversaries, the Army is interrogating how it traditionally does everything from imagery collection, signals intelligence and electronic warfare, hoping to transcend current methods and create a battlefield picture that extends across these geographic divisions. “If you look at how the [National Defense Strategy] describes what we're supposed to do in competition and conflict, we really needed to have the ability to see deep, to look deep to be able to shape the environment for commanders, [and] the ability to sense the environment,” Lt. Gen. Scott Berrier, the Army's director for ISR/G-2, told C4ISRNET in an Aug. 18 interview, adding that after the counterterrorism fight, he did not feel as though the Army's sensors and platforms were in a great place for great power competition. Berrier is departing his post in a few weeks to head the Defense Intelligence Agency, with Maj. Gen. Laura Potter set to pin on a third star and take over as the next G-2 and ISR task force. Under Berrier, the task force has focused on enhancing other Army missions, namely the service's number one modernization priority: long range precision fires. “We really see ourselves as enabling capability ... when you talk about long range precision fires and the sensor to shooter, if you're going to shoot a target at 1,000 miles, you certainly have to see it,” he said. The task force works to corral all the ongoing modernization efforts conducted by Army Futures Command and its various cross functional teams, along with the acquisition community, to ensure they are all coordinated for an integrated, modernized ISR footprint. This means helping to advise on and shape requirements for future systems, while contributing in exercises that test new capabilities and concepts with forces across the world. Other contributors include the Future Vertical Lift and Assured Position, Navigation and Timing teams. The Task Force is also examining to what extent cyber capabilities can play a role in deep sensing, though details are scarce on this front. The Army's Program Executive Office Intelligence, Electronic Warfare and Sensors is contributing through offensive cyber, which officials in the past have said cyber is a collection mechanism. A layered approach The Army's ISR modernization approach begins with the terrestrial or ground layer, Berrier said. The Intelligence Center of Excellence at Fort Huachuca is taking the lead in this arena. The main capability is the forthcoming Terrestrial Layer System-Large, the Army's first brigade-focused, integrated signals intelligence, electronic warfare and cyber platform. Berrier explained that the Army is trying to regain capability it lost after the Cold War. “What we need to have is a sensing platform that can really, really see in the electromagnetic spectrum very complicated signals; to be able to understand [and] perceive the environment; and then — if we want to make an effect inside that environment — [create an effect] with our electronic warfare operators but also ... put an effect into cyberspace,” he said. “We think TLS, with our [brigade combat teams] and those formations, will have what I would call close access, perhaps, to adversary networks. And they'd be able to influence those networks in a number of different ways, as you can imagine.” The Army awarded two prototypes for TLS — to Boeing subsidiary Digital Receiver Technology, Inc. and Lockheed Martin — for a roughly year long experiment with units, after which it will choose one vendor to move forward. There are significant changes for the Army in the aerial layer, namely a new, first-of-its-kind jet the Army is experimenting with. Intelligence and Security Command is heading the aerial layer for the Army currently and just deployed a commercial jet called the Airborne Reconnaissance and Targeting Multi-mission Intelligence System (ARTEMIS), made by Leidos and first reported by Aviation Week, to the Pacific. ARTEMIS is the first step in something the Army is calling the Multidomain Sensing System, which will operate at higher altitudes than the Army has traditionally operated. “Our sensors are flying between 22,000 and 24,000 feet today. We think they need to be much higher ... think in the 40,000 range,” Berrier said. He added there is an unmanned component that could potentially include gliders or balloons. However, he acknowledged the technology might not be ready yet. Ultimately, the Army believes it will need signals intelligence, electronic intelligence, electronic warfare and cyber capabilities in the final Multidomain Sensing System, whatever that ends up being. Berrier described a year long “campaign of learning” for the Multidomain Sensing System, which begins with ARTEMIS in the Pacific. “That will take about a year before I think we're ready to even make a decision. Do we stay sort of in this realm of assets that are around 20,000 feet or 22,000 feet? Or do we, should we go higher to be in that competition ISR fight?” he said, adding the Army will partner with other services on big wing ISR. Finally, the third layer is the space layer, which manifests itself in the Tactical Intelligence Targeting Access Node (TITAN). TITAN is a ground station the Army is building to replace several existing ground stations. Since the Army isn't in the business of building and launching its own satellites, it wants to take advantage of the bevy of satellites already in existence by agencies such as the National Reconnaissance Office. And it believes TITAN will allow it to access these constellations better. Berrier said there will be some processing and artificial intelligence that goes into the system, enabling it to identify targets sooner. The Army is experimenting with TITAN “surrogates” in Europe, through the 66th Military Intelligence Brigade, and in the Pacific through the 500th Military Intelligence Brigade. The Army is also using exercises such as Defender Europe and Defender Pacific along with newer units to include the Multidomain Task Force and its Intelligence, Information, Cyber, Electronic Warfare and Space battalion to help prove out these intelligence concepts and capabilities. It is also working to modernize data standards and dissemination systems such as the Distributed Common Ground System, which is transitioning to the Command Post Computing Environment. Ultimately, Berrier said these ISR modernization efforts are about helping the Army deter conflict. But if that fails, the service needs to be ready for the multidomain battlefield it expects to fight on in the future against near-peer powers. “If you do competition effectively and if you do competition ISR in the right way, you'll never get to conflict because you'll always have a decision or an information advantage over our adversaries,” he said. “If we do transition to conflict, it is about reducing the sensor-to-shooter loop that's going to be so key for multidomain operations. If you want to do MDO ... the ISR Task Force is about bringing multidomain intelligence to competition and conflict.” https://www.c4isrnet.com/battlefield-tech/2020/08/19/the-army-is-working-to-see-across-thousands-of-miles/

• 

  12 novembre 2021 | International, Terrestre, C4ISR

  T-AIS Data Service in Specific Regions