Reproduce “Newton vs the machine: solving the chaotic three-body problem using deep neural networks” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Philip G. Breen, Christopher N. Foley, Tjarda Boekholt, Simon Portegies Zwart
This research from the University of Edinburgh, University of Cambridge, Instituto de Telecomunicações, and Leiden University demonstrates using deep neural networks to solve the chaotic three-body problem. The trained network achieves speedups up to 10^8 times compared to traditional numerical integrators while accurately reproducing chaotic dynamics and maintaining physical conservation via a projection layer.
Key insights
- Deep neural networks can learn the complex, chaotic dynamics of the three-body problem and generalize to unseen initial conditions.
- A post-processing 'projection layer' can effectively enforce physical conservation laws, like energy conservation, on neural network predictions.
- Machine learning offers a viable alternative to traditional numerical integrators for computationally intensive chaotic systems, enabling massive speedups.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.