IU Incubator KIDZ

The primary goal of this research project is the development of powerful and innovative methods and approaches for detecting cyberattacks, particularly zero-day exploits.

By leveraging advanced methods from the fields of machine learning and artificial intelligence (AI), the project will focus on three main aspects: continuous learning, conformal prediction, and explainable AI. Each of these areas addresses specific challenges inherent to cybersecurity, such as the dynamic nature of cyber threats, the need for reliable uncertainty estimation in predictions, and the critical demand for transparency and trust in automated systems.

Each of these research areas not only pushes the boundaries of what is possible with AI/ML in cybersecurity but also addresses the unique challenges associated with securing digital infrastructures against sophisticated and constantly evolving threats. Their integration into cybersecurity solutions is seen as a decisive factor.

The applicants aim to conduct research under the incubator program, focusing on the application and further development of methods from the aforementioned three subfields of AI/machine learning in the context of detecting and mitigating zero-day exploits.

Thus, the project is based on the following two sub-problems (SP) related to zero-day exploits:

• SP1: Delayed detection due to the time lag between the emergence of threats and the adaptation of AI/ML models.

• SP2: Insufficient quantification of uncertainty and lack of transparency in AI/ML models.