About Me:
Hi! I am a doctoral researcher focused on Explainable Multimodal AI at Åbo Akademi University. I completed my master’s degree in information technology through the EDISS master’s program in 2023 before working at a fintech startup as a senior machine learning engineer. My passion for research, along with a strong interest in addressing the ethical and trust challenges posed by powerful technologies like AI, led me to pursue a PhD, which I began in Spring 2024.
Research Focus:
As AI technology advances, we aim to develop models that emulate the human brain’s ability to process multisensory inputs, enabling machines to learn from and respond to diverse information modalities. However, as AI models become more sophisticated, they also become more opaque, making it increasingly challenging to understand and explain their decision-making processes. My primary research interest lies in unraveling how multimodal models interpret and integrate different input modalities, with the goal of creating methods to transparently and effectively explain their decisions in complex settings and systematically quantify them.
Objectives:
- Extensible and interpretable framework: for analyzing various transformer-based multimodal integration techniques.
- Evaluation Benchmark: for cross-algorithm explainable AI (XAI) assessment.
- Validation of the framework: across multiple multimodal domains to provide more accessible results for all stakeholders (users and developers).