Special Session organized by A.V. Tkachenko, S. Maslov and B.M. Mognetti
The rapid progress of artificial intelligence has exposed fundamental limitations of conventional computing architectures, such as energy efficiency, but also serves as an inspiration for learning systems beyond these digital platforms. This led to a growing interest in learning implemented directly in physical and chemical systems, where adaptive behavior and learning emerge from collective dynamics and non-equilibrium interactions rather than explicit algorithms. Recent work shows that disordered materials, metamaterials, active systems, spintronic platforms, and chemical/biomolecular networks can acquire memory, improve performance, and adapt through repeated interactions with their environment. In this perspective, learning becomes a physical process embodied in matter itself. Extending these ideas further, evolutionary dynamics can be viewed as a highly general form of learning, in which populations accumulate information about their environments through variation, selection, and inheritance. By bringing together approaches from statistical physics, soft matter, biochemistry, and evolutionary theory, this session aims to identify unifying principles of learning, adaptation, and emergent intelligence across physical, chemical, and biological systems.
List of potential speakers:
- Menachem Stern (AMOLF, Netherlands) – physical learning in matter
- Vincenzo Vitelli (University of Chicago, USA) – topological mechanics & learning
- Julie Grollier (Université Paris-Saclay, France) – spintronic and neuromorphic physical intelligence
- Andrea J. Liu (University of Pennsylvania, USA) – disordered solids & memory
- Arvind Murugan (University of Chicago, USA) – physical & chemical learning paradigms
- Corentin Coulais (University of Amsterdam, NL) – mechanical metamaterials & memory
- Silke Henkes (Leiden University, NL) – active/adaptive matter
- Ulrich Gerland (Technical University of Munich, DE) – statistical physics of evolution, origin of life
- Stefano Zapperi (University of Milan / CNR, IT) – plasticity & emergent memory
- Michael Elowitz (Caltech, USA) – learning in biomolecular networks implemented via synthetic biology
- David Eric Smith (Earth-Life Science Institute, Tokyo, JP) – nonequilibrium statistical physics, evolution, origin of life