Health Intelligence Lab (HINTLab)

The next step for medical AI is scaling. One of its central bottlenecks is data: not only its scarcity, but also its fragmentation across modalities, institutions, and time. Most medical AI systems are built around individual clinical tasks. We shift the focus from learning tasks to learning patients: building faithful, reusable models of patients from diverse and incomplete observations. If we can learn patients well, many clinical tasks can be solved, too. To pursue this vision, we bring together generative and agentic AI, multimodal learning, and world models to build virtual patients that integrate, reproduce, predict, and interact with clinical observations over time. Ultimately, we aim to extend intelligence from virtual to physical clinical scenarios.

The Bitter Lesson (Richard S. Sutton, 2019) taught us that “general methods that leverage computation are ultimately the most effective.” Yet in healthcare, computation alone is not enough: scaling also requires technical readiness, ethical safeguards, financial feasibility, cultural viability, and sustainable deployment.

We are fortunate to be backed by strong resources: growing public databases, even larger proprietary datasets, established clinical collaborations, and powerful compute through the ELLIS Institute, Aalto University, and Finland's CSC/LUMI supercomputer.

Join us on this journey to shape the future of AI for healthcare and make real-world impact.

Healthcare is vast, and as an academic lab we cannot and should not attempt to cover everything. Rather than chasing a "generalist" dream, we aim to create clinically oriented specialist foundation models that are narrow but deep, and world-leading in their niche. Well-known exemplars from the community include SMPL for human reconstruction and FreeSurfer for neuroimaging.

My understanding of real-world healthcare needs also comes from my past experience as a medical AI entrepreneur, with NMPA-approved (China's FDA) products and multi-million-dollar venture funding. This background shapes the way we think and work at the HINTLab.

- We are a technical lab targeting first-of-its-kind innovation, with the ambition to translate research into real-world impact.

- We are engineers: we solve problems within and beyond the scope of AI.

- We are also scientists: we seek to understand why and emphasize technical elegance.

- We solve real challenges, not just pursue numbers or publications, though we often achieve those too.

- We appreciate clinical collaborations, but we are not satisfied with quick fixes.

- We value hard work, but smart work even more.

- We believe, with Margaret Mead, that “Never doubt that a small group of thoughtful, committed citizens can change the world; indeed, it is the only thing that ever has.”

ELLIS Institute Finland is part of the European Laboratory for Learning and Intelligent Systems (ELLIS), the pan-European network of excellence in AI. The ELLIS Institutes are strategic initiatives of ELLIS to build world-class AI research hubs. As the second ELLIS Institute, the Finnish institute now hosts its first cohort of nine PIs.

Aalto University is a multidisciplinary community in the fields of science, business, and art & design. Established in 2010, it is the highest ranked university in Finland (QS #114) and a vibrant hub for interdisciplinary research, entrepreneurship, and innovation.

Helsinki and Finland consistently rank among the top worldwide in education, equality, safety, and quality of life. Finland offers excellent living conditions and comprehensive social security, whether for individuals or families, and has been repeatedly ranked as the happiest country in the world (World Happiness Report).