Learning Systems Under Attack—Adversarial Attacks, Defenses and Beyond

Deep learning has brought many advances to various fields and enabled applications such as speech and visual recognition to flourish. However, recent findings show that Deep Neural Networks (DNN) still have many problems of their own. The many vulnerabilities present in DNNs unable their application to critical problems. Here, some of these vulnerabilities will be reviewed and many of their possible solutions will be discussed. Regarding legislation, a series of practices will be discussed that could allow for legislation to deal with the increasingly different algorithms available. A small overhead for a safer society. Lastly, as artificial intelligence advances, algorithms should get closer to human beings and legislation itself should face deep philosophical questions in an age in which we will be challenged to reinvent ourselves, as a society and beyond.