You have to catch a flight and you are getting late. In these moments, getting airport security clearance can really frustrate you. And despite this process is slow, the officials do not have a great record at catching threats.
Department of Homeland Security along with the Kaggle Data science community is hosting an online competition to build machine learning powered tools that can augment agents. It should make the entire system more accurate and efficient.
Kaggle was acquired by Google earlier this year. It regularly hosts the online competition where data scientist can compete for more money by developing better approach to more complex machine learning problems.
The competition started by Homeland Security to improve that recognition algorithm will be the Kaggle’s third launch this year. This competition will have more than $1.5 million prize money with the winner getting $500,000. The competitors will have to accurately predict the location of threat object on the body. The TSA is making the data set of images available to competitors so that they can train their program on the people carrying weapons. The images will be the staged image rather than the real-world examples.
There are various other problems that TSA has to face like the expensive physical machines are complicated to upgrade and none have the sophisticated GPU that is used these days.
DHS has promised to work with the winners for developing potential real-world applications. TSA has put a lot of effort in creating the data set as another big issue was the risk of bias influencing the automated threat decision which can be the nightmare to travellers.
Google also plans on making GCP available to competitors in near future and it’s not forcing people to use Tensor Flow (its own open source framework). The competition will close in December. So, if you are interested in AI and ML then you should definitely try in this competition as you never know you might just win it.
Image Source: cbronline.com