Hybrid Digital Twins – The Next Big Breakthrough in AI
Our intent is to make Sweden world-leading in this next big breakthrough in
AI, and use this as glue to build up a new ecosystem for health.
Personalized digital twins
At the heart of this new model for health lies what we believe to be the next big
breakthrough within AI and data-driven life science: personalized digital twins of
individuals, which combines M4 models for how the body works, with machine
learning (ML) and large language models (LLMs).
From intracellular to whole-body
These processes include both the intracellular level, how cells combine to become organs, and how organs combine to the whole-body response. The processes can be simulated while a person is doing these things, but also on a longer timescale: what would happen if the person did this in a certain way for weeks, years, or decades?
M4-model
The name M4-model is short for Multi-level (cell to whole-body), Multi-timescale (seconds to years), Multi-species (animals, microphysiological systems, and humans), and mechanistic models. In other words, these models describe what happens in the body mechanistically, i.e. which processes that happen, when a person is doing something: e.g. eats a meal, exercises, or takes a medication.
World-unique model
We are arguably the only ones in the world that has such a model for all major functionalities in the body (brain activity, metabolism, inflammation, blood flow, etc), and we are also combining these models with ML-models (to describe e.g. the risk of a stroke, or to identify features in large-scale data) and LLMs (which can read scientific papers, and be used to communicate with non-experts).
Integrated knowledge about human
Unlike any of these technologies taken individually, our combined hybrid approach has the potential to integrate all knowledge available about the human body and mind, can personalize this knowledge to a specific person (i.e. create a digital twin for that person), and make that knowledge available to all actors in the new emerging health ecosystem.
The new model for health is a patient-centric ecosystem centered around digital twins, i.e. computer models for patients.
During this initial pre-project (during the spring of 2026), we will set up a plan for how to integrate all sorts of models and data into the twin, do two pilot studies, and expand the eco-system into a full excellence cluster, covering all relevant actors for health. We envision that the full excellence cluster will be world-leading by 2035 in the next big breakthrough in AI for health – hybrid digital twins of individuals - and that it will be big and bold enough to tackle our enormous impending healthcare challenges.