The field of complex systems research is moving towards the development of more sophisticated modeling and analysis techniques. Recent works have focused on improving the accuracy and efficiency of methods for modeling and predicting the behavior of complex systems, including those with hidden states and nonlinear dynamics. Notable advancements include the use of machine learning frameworks for modeling spatio-temporal dynamics and the development of novel methods for signal recovery and uncertainty quantification.
Some noteworthy papers in this area include: The MEET-Sepsis framework, which achieves competitive prediction accuracy for early sepsis prediction using only 20% of the ICU monitoring time required by state-of-the-art methods. The StructureFlow approach, which jointly learns the structure and stochastic population dynamics of physical systems, showcasing its utility for tasks such as structure learning from interventions and dynamical inference of conditional population dynamics. The PDE-Free mass-constrained learning framework, which effectively reconstructs the solution operator of the underlying PDE without discovering the PDE itself, leveraging a manifold-informed objective map that bridges multiple scales.