{"id":28993,"date":"2018-11-12T22:53:33","date_gmt":"2018-11-13T03:53:33","guid":{"rendered":"https:\/\/digital.hbs.edu\/platform-rctom\/submission\/two-hours-prior-bringing-machine-learning-to-the-us-department-of-homeland-security\/"},"modified":"2018-11-12T22:53:33","modified_gmt":"2018-11-13T03:53:33","slug":"hurry-up-and-wait-bringing-machine-learning-to-the-us-department-of-homeland-security","status":"publish","type":"hck-submission","link":"https:\/\/d3.harvard.edu\/platform-rctom\/submission\/hurry-up-and-wait-bringing-machine-learning-to-the-us-department-of-homeland-security\/","title":{"rendered":"Hurry Up and Wait: Bringing Machine Learning to the US Department of Homeland Security"},"content":{"rendered":"

Imagine arriving at a major US airport 30 minutes before boarding time, checking a bag, proceeding through security and arriving at the gate with time to spare for a coffee.\u00a0 This utopia is not beyond the grasp of reality as we explore the potential impact of machine learning on the airport check-in and security processes.<\/p>\n

In 2017 the US airline industry net operating revenues were in excess of $175 Billion [1].\u00a0 With air travel expected to increase in the coming years, this industry is marked by fierce competition.\u00a0 One area in which the major contenders are investing is machine learning with hopes of improving customer service, logistics, and bag checking [2].\u00a0 The competitiveness of this market will undoubtedly drive rapid innovation and implementation, which leaves the government run security screening as the most likely bottleneck in the process of getting from the airport entrance to your aircraft seat.<\/p>\n

Current Security Screening Process<\/strong><\/h3>\n

The Transportation Security Administration (TSA), which is a subset of the Department of Homeland Security (DHS), is responsible for security screening at US airports.\u00a0 At airport security checkpoints, the TSA places several employees along with equipment (x-ray machines, metal detectors, cameras, etc.) to screen each passenger.\u00a0 Figure 1 shows the process flow diagram for a single passenger passing through a TSA security checkpoint.\u00a0 The process depicted begins after the passenger\u2019s ID and boarding pass have been checked.<\/p>\n

\"\"<\/a>
Figure 1: Airport Security Check Point Flow Chart.\u00a0 Source: Journal of Air Transport Management [3].<\/figcaption><\/figure>As denoted by the worker symbols, human interaction is required at multiple points in the process.\u00a0 Many of these decision points, such as ID\/boarding pass check and X-ray screening, are both time consuming and susceptible to human error.<\/p>\n

Progress<\/strong><\/h3>\n

Fortunately, this process is ripe for machine learning intervention, and DHS has taken notice.\u00a0 In 2017 DHS teamed up with Google to introduce a $1.5 million contest to build computer algorithms that can automatically identify concealed items at airport security checkpoints [4].\u00a0 Furthermore, DHS is working with technology companies to develop CT systems that can automatically identify items hidden in luggage while using neural networks to be adaptive in the face of emerging threats [4].<\/p>\n

DHS has even developed an open source web application known as Global Travel Assessment System (GTAS).\u00a0 Developed in response to UN Resolution 2178, GTAS is meant to improve \u201cGlobal Security by using industry-standard Advance Passenger Information (API) to screen commercial air travelers.\u201d [5]\u00a0 Specific data used for algorithm targeting includes:<\/p>\n

\"\"<\/a>
Table 1: API data collected that GTAS uses for targeting. Source: GTAS [5].<\/figcaption><\/figure>Earlier this year, DHS Science and Technology Directorate announced that it had paid $200,000 to DataRobot, Inc. to begin testing a machine learning platform for the US Customs and Border Protection\u2019s GTAS [6].\u00a0 The idea is to simplify the user experience and lower the requirement for a large team of data scientists while still producing increasingly accurate predictive models [6].<\/p>\n

Concerns<\/strong><\/h3>\n

As is the case with all emerging technology that impacts human safety, there are concerns with implementation.\u00a0 As we have seen, the current phase of artificial intelligence is statistically impressive, but often individually unreliable.\u00a0 Machines simply make mistakes that a human never would (recall IBM Watson\u2019s response of \u201cToronto\u201d while answering a question about US Cities on the game show Jeopardy<\/em>).\u00a0 While playing a game, those mistakes are funny, but at an airport security checkpoint the humor falls short.<\/p>\n

Current machine learning technology is also susceptible to \u201ctargeted distortion.\u201d [7]\u00a0 In Figure 2 below, a machine learning system that had been taught to recognize images, correctly identified the image on the left as a panda.\u00a0 However, an engineer with knowledge of the algorithm overlaid the panda image with the center image.\u00a0 The result was the image on the right, which looks the same to the human eye, but caused the algorithm to incorrectly identify the image as a gibbon.<\/p>\n

\"\"<\/a>
Figure 2: Targeted Distortion. Source: DARPA [7].<\/figcaption><\/figure>\n

Moving Forward<\/strong><\/h3>\n

DHS should implement machine learning at airport security checkpoints in order to\u00a0improve the current standard of ID matching and hidden object detection while simultaneously reducing the time it takes passengers to get through the security screening process.\u00a0 Even a small reduction<\/em> in the number of false positives in the passenger and bag screening process would massively reduce the burden on both patrons and staff.<\/p>\n

To reduce public safety risk, DHS should begin implementation on a prescreened population, such as those travelers in TSA Precheck and Global Entry.\u00a0 Human monitoring and quality checks would further reduce the initial risk.\u00a0 As the algorithm adapts and trust is developed, the use case could spread to the entire air traveling population.\u00a0 Long term, DHS should consider applying the machine learning tool to other areas of airport security, such as suspicion behavior recognition and predictive demand algorithms to optimize staff allocation.<\/p>\n

Questions Remain<\/strong><\/h3>\n