Neural Cartographer: Recovering Invisible Laws express
Many natural processes are described by equations, but we only have fragments of them. The new method acts like a cartographer who reconstructs a complete map of the terrain from a few landmarks. Known laws provide a reliable standard framework — like a coastline, while the neural network draws the hidden dynamics by analyzing data. With each cycle, the algorithm refines the map: first it reconstructs the system's full history, then it corrects the unknown parts. This is reminiscent of how astronomers map invisible dark matter in galaxies by its gravitational influence.
Tests showed: the method accurately predicts behavior even when only 5% of all factors are visible. Unlike black boxes, it retains physical meaning, which is crucial for reliable models in medicine and ecology.
Original abstract
as written by the authors arXiv:2607.15180
Ordinary differential equations (ODEs) are widely used to model dynamical systems in physics, biology, neuroscience, and physiology, but in many applications some equations of the dynamics are unknown and only a subset of the state variables are measured. We propose a hybrid neural--physics framework in which the known components of the ODE are kept explicit and the missing components are represented by a neural network. The proposed method consists of two stages where we alternate between state and parameter estimation and iterate until a predetermined criterion is met. Specifically, in the first step, we treat the model parameters as being known and we infer the latent states from the available measurements using a Rauch--Tung--Striebel (RTS) smoother. In the second stage, we treat the smoothed trajectories as being known and use them to estimate the neural networks' parameters through backpropagation. We evaluate the method on benchmark systems spanning linear, nonlinear, and stiff dynamics under partial state observation. Across these settings, the proposed method learns missing ODE components from incomplete measurements while exploiting and retaining interpretable mechanistic structure and improving latent-state reconstruction and long-horizon prediction.
A similar principle is at work in self-driving cars: when sensors momentarily lose sight of an object, a physical model plots its likely path, and a neural network refines the details, just as a cartographer draws a map from a few lines.