580 California St., Suite 400
San Francisco, CA, 94104
Key research themes
1. How do thermodynamic and information-theoretic principles underpin self-organization in living biological systems across scales?
This theme investigates foundational principles explaining how biological self-organization emerges from physical laws, focusing on nonequilibrium thermodynamics, free-energy minimization, and information theory. It addresses how living systems maintain order against entropy, embody environmental models, and achieve multi-scale organization via minimizing variational free energy or entropy, uniting physical, informational, and biological dynamics. The convergence of physicochemical and informational frameworks sheds light on the universal mechanisms underlying life’s complex self-organizing behavior across spatial and temporal scales.
Key finding: This paper establishes that living cells function as disequilibria-converting systems employing information encoded in genetic material to create protein nano-engines that enable self-organization far from thermodynamic... Read more
Key finding: This work extends the free-energy principle (FEP) as a unified explanatory framework for biotic self-organization across scales—from cellular to societal levels—demonstrating how life minimizes variational free energy to... Read more
Key finding: Through simulations informed by variational free energy minimization, this paper introduces Markov blankets as statistical boundaries fundamental to biological self-organization, enabling dynamic separation of internal,... Read more
2. How does confinement influence and steer self-organization across physical and biological systems?
This research area focuses on the role of spatial and environmental constraints—termed confinement—in modulating the emergence and dynamics of self-organized structures and patterns in diverse systems. Confinement alters the degrees of freedom of units, shapes phase space probability distributions, and can act as a catalyst or inhibitor of collective phenomena. Understanding and manipulating confinement enables active control of pattern formation from molecular assemblies to crowds and ecosystems, advancing applications in materials science, biological tissue engineering, and sociotechnical systems.
Key finding: This review synthesizes diverse examples showing that confinement—whether physical boundaries, interfaces, or more abstract constraints—fundamentally alters self-organization by limiting translational and rotational degrees... Read more
Key finding: Introducing entropy by neighbor distance (H_NDist), this paper validates a sensitive, information-theoretic metric to characterize positional order and phase transitions in active matter systems spanning simulated boids,... Read more
3. Can computational and neural-based clustering approaches model and reveal self-organizing principles in data and biological spatial structures?
This theme examines the application of self-organizing maps (SOMs) and related neural algorithms to understand, visualize, and predict complex spatial patterns, such as ecological invasions or biological grouping, revealing emergent structures in high-dimensional data. These methods exploit topological preservation and neighborhood relations to mimic natural self-organization, facilitating insights into how interactions at local levels yield global organization, and enabling the exploration of spatial dynamics and phase transitions in biological and artificial systems.
Key finding: Applying SOMs to presence/absence data of non-native fish species in the River Garonne system, this study successfully identifies distinct habitat clusters associated with invasion patterns, demonstrating SOM's capacity to... Read more
Key finding: This work develops a spherical SOM based on tetrahedral geodesic domes (4HSOM), overcoming visualization and border effects inherent in traditional 2D SOMs and icosahedral spherical SOMs, thereby improving topographic... Read more
Key finding: This paper provides a unified theoretical foundation connecting ant-based clustering algorithms with Kohonen's self-organizing batch maps, demonstrating that the probabilistic picking and dropping behaviors in ant-based... Read more