Discovering sparse interpretable dynamics from partial observations

Nature·Aug 12

✦ NabkaNews BriefAuto-summarized from multiple outlets · verify with the source

Researchers are exploring methods to discover and interpret complex dynamics from incomplete data. Various approaches are being developed, including physics-informed regression and model-based reinforcement learning, to identify governing equations and patterns in fields such as physics and geosciences. These techniques aim to provide efficient and interpretable models of nonlinear dynamics from limited observations.

Full coverage

Moving from table to graph in physics-informed spatio-temporal symbolic regression

Nature·May 23

SINDy-RL for interpretable and efficient model-based reinforcement learning

Nature·Nov 28

Sparse identification of nonlinear dynamics with low-dimensionalized flow representations

Cambridge University Press & Assessment·Sep 6

Automatically discovering ordinary differential equations from data with sparse regression

Nature·Jan 9

Towards data-driven discovery of governing equations in geosciences

Nature·Oct 14

Learning interpretable network dynamics via universal neural symbolic regression

Nature·Jul 6

Physics-informed learning of governing equations from scarce data

Nature·Oct 21

Data-driven discovery of dynamics from time-resolved coherent scattering

Nature·Sep 20

View original report at Nature ↗

Fetching the latest headlines…