Fighting the dark side of AI: threats and defense against data attacks

Attacking datasets poses a significant threat to artificial intelligence (AI) systems, negatively impacting their accuracy and security. In this presentation, we will focus on identifying different types of attacks, such as noise injection, label manipulation, and adversarial attacks, and we will also discuss methods of defense against these threats. The main goal is to raise awareness of the potential risks posed by data attacks in order to ensure the integrity and reliability of AI-based systems.