Applied Complex Systems Science: Workflows for Efficiently Matching Models and Experiments

Driven by several converging developments, the relationship between modelling and experimentation in complex systems has entered a new phase. The abundance of high-resolution data, advances in computational power, and the integration of artificial intelligence have created a growing demand for phenomenological, rapid-response modelling frameworks. These "in silico" environments are no longer just theoretical exercises; they are tailored to the requirements of specific experimental contexts, such as multi-scale models of biophysical and bio-inspired systems, active matter, model-supported therapies (ranging from cellular therapies to physio- and occupational therapy), and socio-technological systems. This evolution calls for orchestrated workflows where model variables are explicitly mapped to experimental observables, ensuring that theoretical abstractions remain grounded in measurable physical quantities. Importantly, these measurable quantities are determined not only by the system but also by the available experimental facilities. This, in turn, necessitates early communication and suitable workflows between modellers and experimentalists.In traditional physics, models often represent general solutions where the relationship between measured observables and variables is close to a bijection. In such cases, models serve primarily to test and refine our fundamental understanding. However, as we move toward complex physical systems—characterised by emergent behaviour, multi-scale interactions, and non-equilibrium states—this mapping becomes obscured.In these complex domains, measurements are often indirect, consisting of datasets of a dimensionality that depends on the experiment and can be high or, as it often occurs, is quite low, or spatio-temporal patterns (e.g. fluorescence intensities caused by complex chemical dynamics involving several reaction pathways). A model representing this data relies on the underlying governing equations, which are not always immediately apparent. Here, models play a dual role: they must advance theoretical frameworks (such as system-specific networks of interactions or phenomenological dynamics) while simultaneously serving as essential tools for data interpretation and state estimation. To be effective, these models must be contextualised to the specific experimental setup, bridging the gap between microscopic interactions, (potentially mesoscopic) emergent dynamics and macroscopic observations. Achieving this alignment raises profound conceptual questions regarding coarse-graining and representation. For instance, how do specific data-acquisition frequencies or spatial resolutions bias the resulting physical model? A common tension exists between the aspiration for "first-principles" generality and the practical efficiency of "effective theories" tailored to specific applications. Deepening our understanding of this balance—and formulating the methodological principles to guide it—is crucial for the next generation of physical sciences.Contributions to this Research Topic are invited to address one or more of the following themes:• Conceptual and epistemological discussions on the role of models and effective theories in complex physical, biological, active, or socio-economic systems.• Hybrid modeling approaches that combine mechanistic and data-driven methods, including machine learning, particularly in experimentally constrained or real-time settings.• Best practices, methodological frameworks, and interdisciplinary workflows that enable robust alignment between models and experiments in high-dimensional and multiscale systems even if there is a gap between the dimensionality of the system and the measured data.• Case studies that translate modeling strategies from idealised academic systems to noisy real-world, industrial, or societal applications.• Communication of modeling results to non-specialist audiences, including strategies for conveying uncertainty, scale dependence, and model limitations. Topic editors: Rudolf Marcel Füchslin, Martina Rebekka Spiess, Martin Hanczyc, Mathias Sebastian Weyland, Oliver Castell Source: https://www.frontiersin.org/research-topics/78545/applied-complex-systems-science-workflows-for-efficiently-matching-models-and-experiments