Meet Veronica and Qing, our new Associate Professors

We happily welcome Veronica Lachi and Qing Liu as new Associate Professors at SFI Visual Intelligence.

By Petter Bjørklund, Communications Officer at SFI Visual Intelligence

Both Liu and Lachi recently started their positions at SFI Visual Intelligence's hub in Tromsø, Norway.

Graph-based models applied to real-world domains

Lachi is originally from Italy, and has a background in Mathematics. She obtained her PhD in Artificial Intelligence from the University of Seina, which focused on machine learning and AI methods.

Her research focuses on graphs and graph neural networks. She mentions that these are particularly interesting, because real-world data such as physical systems, sensor networks, biological structures, and relational data, can naturally be modelled as graphs.

"Applying deep learning to these structures enables more realistic modeling, which can lead to improved performance across a wide range of applications," Lachi explains.

She is especially interested in applying graph-based models to domains like energy systems and medical applications, with a focus on uncertainty estimation and explainability.

She will mainly contribute to Visual Intelligence's "Medicine and health", "Earth observation", and "Energy" domains, as graph-based models can play a key role in modeling complex structured data.

"I am very much looking forward to applying my expertise in graph neural networks to real-world, high-impact problems within Visual Intelligence. In particular, I am excited about collaborating across disciplines and contributing to applications in areas such as energy, earth observation, health, and medicine, where robust, well-calibrated, and interpretable AI models are essential," Lachi says.

Medical foundation models and multi-agent AI systems

Liu is from China, and holds a PhD in Computer Science from Central South University, China. Before joining the Visual Intelligence, she was an Academy Research Fellow at the Center for Machine Vision and Signal Analysis, University of Oulu, Finland, for three years, where she led a research group focusing on AI for healthcare.

Liu's research interests lie in machine learning with imperfect clinical data, including self-supervised learning, partially supervised learning, and incomplete multimodal learning. Her research aims to address real-world clinical challenges where labels or modalities are incomplete or unavailable, and to develop intelligent healthcare systems that make healthcare services more available, accessible, and affordable for broader populations.

Qing has shown strong enthusiasm for medical foundation models and multi-agent AI systems. She aims to build a multimodal, multi-agent system using self-supervised learning, partially supervised learning, and reinforcement learning methods.

"This system will be designed to integrate and analyze diverse clinical data modalities including but not limited to clinical notes, color fundus images, OCT scans, and Scanning Laser Ophthalmoscopy images to support comprehensive, personalized eye disease management, spanning preliminary screening, secondary diagnosis, and long-term follow-up," Liu says.

As an Associate Professor at Visual Intelligence, she will also closely collaborate with clinical partners, industry, and interdisciplinary researchers to build the multi-agent eye healthcare system she envisions and to translate advances in artificial intelligence into real-world healthcare applications.