Webbläsaren som du använder stöds inte av denna webbplats. Alla versioner av Internet Explorer stöds inte längre, av oss eller Microsoft (läs mer här: * https://www.microsoft.com/en-us/microsoft-365/windows/end-of-ie-support).

Var god och använd en modern webbläsare för att ta del av denna webbplats, som t.ex. nyaste versioner av Edge, Chrome, Firefox eller Safari osv.

Anahita Baninajjar: Att förstå och förbättra pålitlig AI

Namn: Anahita Baninajjar
Datum för disputation: 11 juni 2026
Titel på avhandling: Towards Trustworthy Machine Learning in High-Stakes Decision Making

Profile picture of Anahita Baninajjar
Anahita Baninajjar

My PhD focuses on understanding and improving the trustworthiness of machine learning models in high-stakes applications like healthcare. Modern AI systems can achieve impressive accuracy, but they can still fail in unexpected ways, leak sensitive information, or behave unfairly. During my PhD, I developed mathematical methods to analyze these models and understand when they are truly reliable. Ultimately, the goal of my work is to help make AI safer for real-world decision-making.

What made you want to pursue a PhD?

I’ve always enjoyed learning, especially mathematics, and over time I realized that I was also drawn to research and problem solving. What appealed to me most was the opportunity to explore difficult questions in depth and develop new ideas and solutions. A PhD felt like the natural path to explore those kinds of problems more deeply.

What is the most fascinating or interesting with your thesis subject?

What I find most interesting is that high accuracy alone tells us very little about whether an AI model is actually reliable. A model can perform extremely well and still fail in unexpected ways, which raises many challenging questions about trustworthiness.

Länk till avhandlingen (på engelska, Lunds universitets forskningsportal).