At the LACSII we conduct research on cybersecurity from the perspective of recent developments of Artificial Intelligence (AI).

Data has become the main feedstock of tools such as machine learning (ML) and data science. ML models are often trained with a massive amount of centralized data available on a server. This exposes individuals to privacy risks and organizations to legal problems if data is not protected properly. To mitigate that and promote data privacy, we focus on the decentralized learning paradigm, namely Federated Learning, where machine learning problems are solved collaboratively on edge devices, under the supervision of a central server or service provider.

The second axis of our research is biometrics robustness. We investigate the countermeasures of voice biometrics towards a variety of voice spoofing attacks. A particular concern is the dissemination of misinformation that can culminate in confusion and risk-taking behaviors. This is aggravated by deep fake signals and the expansion of social media.