Research

LEARNING in LIVING NETWORKS

The emergence of life hinges on information processing: organisms have to interact with the world to find food and mates and avoid traps and predators. To coordinate information processing, highly complex neural networks have evolved. With its 80 billion neurons, the human brain is arguably one of our most complex organs. Importantly, the living neural networks can learn in a local, unsupervised manner; thus without an explicit teacher signal they adapt their connection weights between pairs of neurons autonomously. It is at the core of our research to understand the basic principles of such emergent information processing.

Likewise, information also spreads in social networks. Interestingly, the self-organization in these social networks bears some resemblance with that in the brain if one takes a statistical physics perspective: Both feature non-reciprocal, non-local, and non-static (i.e. learning, adaptive) interactions. Thus, both present prime non-equilibrium systems, and in both, information (and misinformation) spreads largely in a decentralized, emergent manner. By studying neural and social networks side-by-side, we carve out the basic principles that govern their function and dysfunction. Even more, by abstracting from their specific details, our long-term goal is to develop a general theory of living adaptive networks.

You find our publications on Google Scholar or arXiv.