Plus: ICE deploys secretive phone surveillance tech, officials warn of Chinese surveillance tools in US highway infrastructure, and more. Right-wing internet personality and Turning Point USA cofounder Charlie Kirk was shot and killed on Wednesday during a speaking engagement at Utah Valley University in Orem, Utah. After a chaotic 24-hour manhunt, the FBI named 22-year-old Utah resident Tyler Robinson as a suspect in the murder. As polarization and political violence continues to increase in the US, a new platform from the Public Service Alliance is offering tools like data-removal services and threat monitoring to public servants who increasingly need to defend themselves and their data. Meanwhile, new research this week warned that the number of US investors putting money into invasive commercial spyware rose significantly in 2024.

As part of Department of Transportation's plans to modernize the U.S.'s highway infrastructure, the agency issued a new contract opportunity seeking artificial intelligence and data analytics-based solutions. A program under the agency's Federal Highway Administration called for proposals featuring artificial intelligence technology to improve the current national highway design system. The technology developed within this contract will impact the planning, construction and maintenance of the nation's highways. A spokesperson for the FHWA informed Nextgov the projects' goal is to successfully transform highway transportation with AI technology, and that, following the projects' individual outcomes, officials will decide how to implement it across the transportation sector. "With the growing number of maturing and commercial applications, there still is a need for early state research to support emerging advances in AI that can solve even more complex questions in highway transportation," the RFI notes.

This chapter focuses on Deep Learning and techniques that can be used to keep neural networks from getting out of hand as their complexities get deeper. Traditionally Deep Learning is defined as a neural network that contains 3 or more layers. But, with this addition of layers comes additional complexity and with complexity comes more ways for a project to break. Most of this chapter deals with introducing us to the techniques that we can use to minimize these breakages when training deep models. Neural Networks are trained through backpropagation using gradient descent to adjust their weighting so that we get the intended result.