News

By: Amber Corrin The Navy released a long-awaited final request for proposals Sept. 18 for the re-compete of its Next Generation Enterprise Network contract. But it's part one of two, covering only the hardware side of things as the service looks to overhaul its Navy-Marine Corps Intranet. According to analysts at Deltek, each piece of the NGEN-R request is valued at roughly $250 million over a three-year period, per estimates from Space and Naval Warfare Systems Command. That's significantly lower than NGEN's original $3.5 billion price tag. Specifically, the RFP seeks hardware devices for use on the Department of Defense's classified and unclassified networks, including desktops, laptops, two-in-one detachable devices, tablets, ultra-small desktop computers, as well as thin- or zero-client devices. A single device could serve multiple users and associated accounts, according to the RFP. But for the roughly 400,000 devices NGEN-R looks to replace, the service in particular is looking at an end-user hardware-as-a-service arrangement. “It's breaking out the services that are being provided in a way that allows us to gain most effective advantage of how industry does business today,” Capt. Don Harder, deputy program executive officer for Navy enterprise information systems, told Federal Times in a recent interview. “The end user of hardware and devices as its own separate contract, there are those suppliers out there that that's what they specialize in. By breaking that out into its own contractual component within the NGEN-R construct ... we believe will allow us to get more effective advantage to pricing on those components.” The language in the RFP solidifies Harder's thoughts as part of the statement of work. “In acquiring EUHWaaS, the Government is only acquiring the service of using an EUHW device. This is not a purchase, and titles for all EUHWaaS devices remain with the Contractor,” the RFP states. “EUHWaaS includes the provisioning, storage of spares, configuration, testing, integration, installation, operation, maintenance, [end-of-life] disposal of NIPRNet and SIPRNet EUHW, and internal storage device removal and destruction requirements.” Bids for the hardware piece of NGEN-R are due Nov. 19. The second part of the NGEN-R RFP, service management integration and transport or SMIT, is expected in the next 30 days, according to a Navy spokesman. SMIT will cover much of NMCI's backbone and functionality, including services ranging from help desk to productivity suites to network defense — and how they're technically provided. Splitting NGEN-R into two separate contracts was an intentional move designed, at least in part, to give the Navy greater flexibility in the capabilities available to users, and the options for buying them, as technology evolves. “We are modifying how the services are broken out in a way that it allows us to sever some of those services as new mechanisms [and] provide [them as they are] brought into play or brought to our attention,” Harder said, using cloud capabilities as an example. “We may allow a mechanism to pull some of those into either a hybrid cloud or a cloud solution in the future. If so, it may go on a separate contractual vehicle at which point in time we would sever those services away from the SMIT vehicle. So, we're looking at how we take those services and how we manage them contractually, which would allow us, again additional flexibility later on down the road.” Harder said that throughout the development of NGEN-R, he's been eyeing not just the Navy, but also the broader government to benefit from the new approach. “We're building in that flexibility that allows the government the ability in the future even to find components of services that can be done in a more effective or efficient way [and] either sever them or modify them separately as opposed to having to break apart the entire contract to do something,” he said. The hardware piece of NGEN-R was released less than two weeks after Navy officials announced a one-year, $787 million extension to the incumbent provider, Perspecta. Harder declined to put a dollar figure on the NGEN-R contract, as did other Navy officials. The RFP comes after several delays — officials previously had said the contract would be up for bidding this summer. According to Harder, prior to release the RFP had to be approved by leadership at the Office of the Assistant Secretary of the Navy for research, development and acquisition, as well as the Office of the Secretary of Defense's Defense Procurement and Acquisition Policy office. Harder said the Navy has taken extra time to shore up “the education piece” — ensuring the contracting process meets leaders' expectations, particularly with the new strategy. And IT modernization also has come into play, with officials from the broader DoD looking to NGEN as a possible model or even contract vehicle for defense networks down the line, he said. “We need to ensure that what we have placed in the contract and how we're going about the contract meets leadership expectations. And because we are doing things in a different way, that's taking a little bit of time,” Harder said. The Navy's approach to running NMCI today is “one of the more cost-effective ways of managing networks. And there is a desire as part of one of the many IT reform efforts [for possible] integration of networks in the future to mimic or, potentially, even ride on our contracts.” https://www.federaltimes.com/acquisition/2018/09/19/ngen-r-what-is-the-navy-thinking

John M. Donnelly The Pentagon will soon have received about $2.3 billion in the last nine years — money the military never requested — for a special fund intended to help replace earmarks after Congress banned them, our analysis shows. Buried deep inside the $674.4 billion Defense spending measure for fiscal 2019 that the Senate is expected to vote on this week is a chart with one line showing a $250 million appropriation for the Defense Rapid Innovation Fund, the latest installment of sizable funding for a largely unknown program that quietly disburses scores of contracts every year. To supporters, the fund is a way to bankroll innovative systems that the military may not yet know it needs. To critics, the fund is just earmarking by another name. The kinds of systems that net contracts from the innovation fund run the gamut. In fiscal 2016, they included programs to demonstrate artificial intelligence systems for aerial drones, anti-lock brakes for Humvees and underwater communications systems for undersea drones. The systems may be technologies for which the military services have not yet established a requirement because they may not know what is technically possible. It is not clear how many of the systems actually become operational. The defense fund's eclipsing of the $2 billion mark comes as debate heats up in Washington over whether to revive earmarks. And the special account highlights key elements of that debate. Talk of earmarks 2.0 Earmarks have generally been defined as parochial spending, directed by lawmakers and received by people who have not competed for it. In 2011, after earmarks were tied to several scandals and spending projects seen as excess, Congress barred them — or at least a narrow definition of them, critics contend, noting that, among other loopholes, committees could still add money for parochial projects without spelling out who supports them. President Donald Trump suggested earlier this year that a return of earmarks, which were often used in horsetrading for votes, might be beneficial. Minority Whip Steny H. Hoyer of Maryland, has suggested he would aim to bring back earmarks if his party takes control of the House next year. The senior Democrat on Senate Appropriations, Patrick J. Leahy of Vermont, has also supported a comeback for the practice. Republican leaders are less vocal right now, but many of them also support a return to earmarks. “I don't doubt that the next organizing conference for the next Congress will probably wrestle with this issue,” outgoing House Speaker Paul D. Ryan told reporters earlier this month. Account quietly amasses funds The Defense Rapid Innovation Fund was launched in 2010 (first as the Rapid Innovation Program) in the fiscal 2011 defense authorization law. It was a way to capture what proponents called the innovative spirit of programs called earmarks that were clearly about to be banned. Unlike earmarks, the defense fund's money would be competitively awarded by the Pentagon, not directed by Congress, supporters of the idea pointed out. Democrat Norm Dicks, then a senior Defense appropriator, and other advocates of the program described it at the time as a way to capture the innovation among smaller companies, including many who had received earmarks. “We have not always had an adequate way of bringing these smaller firms and their innovation into the defense pipeline,” Dicks said in 2010. Each year since its creation, the fund has received another installment of funds, never less than $175 million or more than $439 million. The program has awarded several hundred contracts, averaging about $2 million each, mostly for small businesses with technologies that were relatively mature and that could address some military need, according to a fiscal 2017 Pentagon summary of the program's results. Full article: http://www.rollcall.com/news/politics/obscure-pentagon-fund-nets-2-billion

By: Mark Pomerleau In its first formal cyber strategy document in three years, the Department of Defense said it would focus its cyber efforts on China and Russia and use the Pentagon's cyber capabilities to collect intelligence as well as to prepare for future conflicts. According to an unclassified summary and fact sheet released Sept. 18, the documents lay out a vision for addressing cyber threats and addresses the priorities of the department's National Security Strategy and National Defense Strategy, which focused on a new era of strategic great power competition. “The United States cannot afford inaction,” the summary reads. It notes that China and Russia are conducting persistent campaigns in cyberspace that pose long term risk. The documents also say that China is eroding the U.S. military's ability to overmatch opponents and that Russia is using cyber-enabled information operations to influence the U.S. population and challenge democratic processes. The DoD's strategy comes on the heels of other major movements in cyberspace from the department. These include the elevation of U.S. Cyber Command to a full unified combatant command — which affords new and exquisite authorities — the full staffing of Cyber Command's cyber teams, an update to DoD's cyber doctrine and new authorities delegating certain responsibilities from the president to DoD to conduct cyber operations abroad. The summary's lists five objectives for DoD's cyberspace strategy: - Ensuring the joint force can achieve its missions in a contested cyberspace environment; - Strengthening the joint force by conducting cyberspace operations that enhance U.S. military advantages; - Defending U.S. critical infrastructure from malicious cyber activity that alone, or as part of a campaign, could cause a significant cyber incident; - Securing DoD information and systems against malicious cyber activity, including DoD information on non-DoD-owned networks; and - Expanding DoD cyber cooperation with interagency, industry, and international partners. The strategy also describes the need to remain consistently engaged with this persistent adversary and to “defend forward” as a means of disrupting or halting malicious cyber activity at its source, including activity that falls below the level of armed conflict. While academics have criticized the U.S. response to Russian election interference, the strategy notes that the United States tends to view conflicts through the binary lens of war or peace while competitors such as Russia see themselves constantly engaged in a state of war. U.S. Cyber Command's new leader is taking a different tact. “We've got to act forward outside of our boundaries, something that we do very, very well at Cyber Command in terms of getting into our adversary's networks. That's this idea of persistent engagement, the idea that the adversary never rests, so why would we ever rest,” Gen. Paul Nakasone said during an August dinner hosted by the Intelligence and National Security Alliance. Nakasone also has described the notion of defending forward as enabling forces to act outside the boundaries of the U.S. to understand what adversaries are doing in order to better defend against them. https://www.fifthdomain.com/dod/2018/09/19/department-of-defense-unveils-new-cyber-strategy

By: Ryan Hazlett With unmanned systems becoming ever more ubiquitous on the battlefield, the question of where unmanned systems and accompanying technologies, such as autonomy, are headed is in the limelight. First, to better understand the future direction of the unmanned field, it is instructive to note some important trends. The number of uses for unmanned systems on the battlefield has increased significantly in the post-9/11 conflicts in Afghanistan and Iraq, with the U.S. Army's Shadow® Tactical Unmanned Aircraft System (UAS) program having logged nearly 1 million flight hours in those areas of operation. The proliferation and commoditization of UAS capabilities is a global phenomenon, as demonstrated by both the widespread possession of UAS hardware as well as the ability to indigenously produce at least rudimentary unmanned systems. Growth of the nascent commercial unmanned systems market has added to this trend, as has the government's emphasis on a greater use of commercial off-the-shelf solutions. But while commoditization has occurred at the platform level — particularly among smaller airborne vehicles — overcoming the challenges of adversaries employing anti-access area-denial (A2AD) military strategies requires far more capable solutions than simply having hordes of cheap drones. In this environment, how will U.S. and allied forces retain their advantage? Critical capabilities and technologies are necessary. These include the ability to dynamically swarm, conduct automatic target recognition, possess on-board autonomy and artificial intelligence, as well as have interoperable communications capabilities. First, future platforms — manned or unmanned — will increasingly need better collaboration between the sensors and payloads they carry and with allied forces. This growing level of collaboration and autonomy is already happening. Driven by advances in onboard computing power, as well as smaller and less power-intensive sensors and advanced algorithms, tomorrow's unmanned systems will be able to better communicate among themselves and make their own decisions on basic functions, such as navigation, to enable dynamic swarming or to identify areas of interest during intelligence, surveillance and reconnaissance missions. Next, systems that can seamlessly operate and communicate with other military platforms across domains will be the most successful. Gone are the days when largely mission-specific platforms dominated the force composition. With platforms needing to be highly capable to meet A2AD threats, a mission-specific approach will simply be unaffordable. Instead, increasingly we see platforms that can act as highly capable but also flexible “trucks” that can easily swap payloads designed for specific missions, while the overall platform serves many needs. Multi-domain abilities for conducting command and control (C2) and other tasks will also be vital as technologies move from remote-control type operations to more of a “man monitoring the loop” concept. Technological progress in providing secure communications and a level of onboard artificial intelligence are necessary enablers, as will be data fusion technologies. Initial versions of these multi-domain C2 solutions for unmanned systems are already here. For example, the U.S. Army has years of experience operating the Universal Ground Control Station and One System Remote Video Terminal that allow soldiers in tactical units to access overhead sensor video from unmanned aircraft. Next-generation, multi-domain control and collaboration technologies to take the concept to a new level are mature, allowing a single user to simultaneously operate multiple vehicles and sensors, including the ability to control numerous types of aircraft and other multi-domain unmanned systems from different manufacturers. In addition, these systems are ready to incorporate the best available software applications as “plug-ins” to an open architecture. Industry is also investing in additional technology to ensure that tomorrow's unmanned systems continue to meet U.S. and allied needs. Among them are advanced power generation, systems with improved maneuverability, and vehicles designed to deploy with lighter support and operational footprints. Done smartly, the application of technologies such as autonomy can be better integrated into unmanned systems to enable improved navigation, intelligence, surveillance and reconnaissance, as well as other tasks, while leaving a man in the loop for the use of weapons. Moreover, defense users can rightly leverage the commercial sector's work on areas such as self-driving cars and unmanned taxis that are at the forefront of artificial intelligence for navigation. But while the military can leverage such commercial developments, there are, and will remain, cyber hardening, survivability and other specific requirements that are unique to the defense marketplace and require experienced industrial partners with deep knowledge of national security needs. The ongoing move away from only long-term programs of record to the embrace of the “buy, try, and decide” model, as well as greater uses of funded prototyping, is helping to fast-track many of these promising new technologies. Companies can now match their internal research and development funding to move that innovation along and ensure the United States and its allies remain at the forefront of unmanned technologies. Ryan Hazlett is senior vice president at Textron Systems. https://www.c4isrnet.com/thought-leadership/2018/09/19/the-new-critical-capabilities-for-unmanned-systems

Par Alain Barluet Si les coups de boutoir de Donald Trump contre l'Otan ont provoqué une prise de conscience importante, les Européens ne parviennent toujours pas à structurer un projet commun. Certains chiffres parlent d'eux-mêmes: moins de la moitié des chars en service dans les armées de l'UE sont de conception européenne et 20 % seulement pour l'artillerie. La propension limitée des Européens à «acheter européen» pour doter leurs forces, la grande disparité des matériels qu'ils utilisent (60 types d'équipements terrestres différents dans l'Union, contre 20 aux États-Unis) illustrent le chemin qui reste à parcourir sur le chemin d'une Europe de la défense. Et encore ne s'agit-il là que du domaine capacitaire. Pourtant, depuis l'an dernier, les conditions d'une prise de conscience ont progressé. Les coups de boutoir du président américain contre l'Otan, qu'il juge «obsolète», et les Européens, qu'il considère comme trop peu investis dans leur défense, ont provoqué une onde de choc de ce côté-ci de l'Atlantique. Un certain nombre de pays, dont la France, ont augmenté leur budget de ... Article complet: http://www.lefigaro.fr/international/2018/09/19/01003-20180919ARTFIG00267-l-europe-de-la-defense-n-avance-qu-a-petits-pas.php

By: Leo Shane III and Joe Gould WASHINGTON — With the fiscal year winding down, Senate lawmakers on Tuesday advanced a multi-agency appropriations deal that would prevent a government shutdown and give the Defense Department its full-year budget on schedule for the first time in a decade. The measure, which provides for more than $606 billion in base defense spending and nearly $68 billion more in overseas contingency funds, is in line with White House requests and spending targets outlined in the annual defense authorization bill approved earlier this summer. “After subjecting America's all-volunteer armed forces to years of belt tightening, this legislation will build on our recent progress in rebuilding the readiness of our military and investing more in the men and women who wear the uniform,” Senate Majority Leader Mitch McConnell, R-Ky., said before the Senate vote. The funding total — approved by a 93-7 vote — amounts to an increase of more than 3 percent for military spending in fiscal 2019, but as important as the boost is the timing of the measure. In recent years, Congress has struggled to pass any appropriations measures before the start of the new fiscal year, relying instead on a series of budget extensions to avoid partial government shutdowns. That has infuriated Pentagon leaders, who have said the fractured appropriations process prevents them from keeping equipment purchases and new program starts on time. If the House finalizes the appropriations measure next week and President Donald Trump signs it into law in the following week (all parties involved have already signaled they expect to do so ), it will mark the first time since 2008 that Congress and the White House have passed their spending plans on time. Senate Armed Services Committee Chairman Jim Inhofe, R-Okla., called that “a major victory” for Congress and the military. The measure funds a 2.6 percent pay raise for troops starting next January and a boost in military end strength of 16,400 spread across the active-duty and reserve forces. Operation and maintenance spending totals $243.2 billion of the defense total, and research and development efforts another $96.1 billion. Defense health and military family programs would receive $34.4 billion. The appropriations fund 13 new Navy ships ― including three DDG-51 guided missile destroyers and two Virginia-class submarines ― 93 F-35 aircraft, 58 UH-60 Black Hawk helicopters, 66 AH-64 Apache helicopters, 13 V-22 aircraft, and $1.5 billion for the upgrade of 135 Abrams tanks. The National Guard and Reserve Equipment Account would also see a $1.3 billion boost from the appropriations plan. In order to avoid political fights over non-defense spending levels, lawmakers agreed to package the military budget bill with the full-year funding for the Departments of Health and Human Services, Labor, and Education. In addition, the legislation contains a budget extension for a handful of agency budgets lawmakers have not yet finished negotiating. The move will prevent a government shutdown at the end of the month, when the fiscal year ends. Several senators lamented before the vote that all of the appropriations bills have not yet been finalized, but for the first time in years, defense advocates aren't among those complaining. In addition to the full Defense Department appropriations plan, lawmakers last week finalized a spending plan for military construction projects and the Department of Veterans Affairs, covering nearly all aspects of national defense and military personnel spending. https://www.militarytimes.com/news/2018/09/18/as-deadline-approaches-senate-advances-674-billion-defense-budget-bill

By SYDNEY J. FREEDBERG JR. “We're looking at airframes of the future that will have common cockpits, advanced propulsion systems, (and) signature management," Miller said. The goal “really is understanding (how) to modify or build an airplane that allows us to operate through that threat environment." AFA: Air Mobility Command's tankers and transports would be big, slow targets in a major war, but without them, most of the US military can't move. The imperative to fly fuel, supplies, and troops in the face of high-tech threats – from anti-aircraft missiles to cyber attack – is forcing AMC to change its approach to aircraft upgrades, communications networks, and what they ask airmen to think about every day, its new commander told reporters here this morning. AMC wants to stimulate innovative thinking by all its people, Gen. Maryanne Miller said, but “not so much on innovation for innovation's sake” – they have to be “much more focused.” On what? “It needs to be on our resilient and agile response,” she said, “being able to operate in that contested, degraded, or operationally challenged threat environment.” There are a lot of buzzwords in that sentence, but they add up to a major change in mindset for strategic transport and logistics. While roadside bombs have ravaged ground convoys in Afghanistan and Iraq, US pilots and sailors can usually assume that they'll arrive alive. Until last year, the joint Transportation Command – overseeing Air Mobility Command, Military Sealift Command, and the Army's Surface Deployment and Distribution Command – didn't even factor into its war plans that an enemy might shoot down planes or sink ships. But Air Mobility Command is already under constant attack in cyberspace, and advanced adversaries such as Russia, China, or even Iran have long-range missiles to challenge US dominance of the air and sea – what's known as an Anti-Access/Area Denial (A2/AD) strategy. Transportation Command is now well into a congressionally-directed Mobility Capabilities & Requirements Study (MCRS) that “is looking at all the things that you're describing,” Miller said when I asked about such factors. TRANSCOM is working with the Pentagon's independent office of Cost Assessment & Program Evaluation (CAPE) and the Air Force, she said, with AMC providing extensive data on how it's operated over the years. Due out “later this fall,” she said, the study will look at the evolving threats and make recommendations on how many tankers and transports AMC needs, and of what kinds. Full article: https://breakingdefense.com/2018/09/how-air-force-tankers-transports-can-survive-in-high-tech-war

By: Brandon Knapp The Defense Intelligence Agency (DIA) awarded spots on a contract worth up to $500 million to eight defense companies and one research organization as part of the agency's effort to buy research, development, technical, and engineering services for intelligence missions, according to an announcement from the Department of Defense Sept. 14. DIA selected AT&T, Booz Allen Hamilton, Harris Corp., KeyW Corp., Leidos Inc., Lockheed Martin Corp., Macaulay-Brown Inc., Northrop Grumman Corp., and Southwest Research Institute for the procurement program, known as HELIOS. The companies won five year contracts that include five additional one-year options. The award winners will have multiple opportunities to compete for task orders on the indefinite delivery / indefinite quantity contract in support of DIA Directorate for Science & Technology missions. Unclassified documents released in 2017 describe in part the scope of HELIOS as offering support services for “sensor technology focus areas including Measurement and Signature Intelligence” as well as “cross-cutting technology areas such as unattended remote sensing, robotics...and data processing, exploitation and dissemination.” Mission areas for HELIOS include space situational awareness, foreign counter space threats, the Internet of Things, data exfiltration, and rapid prototyping. The Virginia Contracting Activity received nine responses to their original open solicitation posted in Dec. 2016. All nine solicitors received a contract with a minimum award of $10,000. The HELIOS program has an estimated completion date of September 2028. https://www.c4isrnet.com/intel-geoint/2018/09/18/dia-announces-winners-in-massive-intelligence-technology-contract

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