Model Discovery

Model discovery focuses on automatically identifying mathematical models that accurately describe complex systems from data, aiming for both predictive accuracy and human interpretability. Current research emphasizes sparse regression techniques, including novel algorithms like SPRINT, and the use of neural networks, such as Physics-Informed Neural Networks (PINNs) and constitutive artificial neural networks, often incorporating domain knowledge and constraints to improve model reliability and interpretability. These advancements are significant because they accelerate scientific understanding by automating the process of model building, enabling efficient analysis of large datasets and potentially leading to breakthroughs in diverse fields like materials science and ecological modeling.