Newton vs the machine: solving the chaotic three-body problem using deep neural networks · HackerPapers discussions

Newton vs the machine: solving the chaotic three-body problem using deep neural networks — HN submission

Tue, 28 Jul 2020 12:09:54 GMT

Solving the chaotic three-body problem using deep neural networks (2019)

46 points; 33 discussion replies on HN. Paper and source history. A submission is not an endorsement.

Newton vs the machine: solving the chaotic three-body problem using deep neural networks — HN submission

Sun, 10 Nov 2019 13:14:40 GMT

Newton vs. the machine: solving the 3-body problem using deep neural networks

1 point; 0 discussion replies on HN. Paper and source history. A submission is not an endorsement.

Newton vs the machine: solving the chaotic three-body problem using deep neural networks — HN submission

Wed, 06 Nov 2019 00:05:24 GMT

Newton vs. machine:solving chaotic three-body problem using deep neural networks

2 points; 0 discussion replies on HN. Paper and source history. A submission is not an endorsement.

Newton vs the machine: solving the chaotic three-body problem using deep neural networks — HN submission

Sun, 27 Oct 2019 09:33:23 GMT

Solving the chaotic three-body problem using deep neural networks

2 points; 0 discussion replies on HN. Paper and source history. A submission is not an endorsement.

Newton vs the machine: solving the chaotic three-body problem using deep neural networks — mentioned by jedbrown

Sat, 26 Oct 2019 21:05:15 GMT

The initial conditions are all stationary, and it's in the plane exploiting symmetry, so they are just 2 real numbers. They don't evolve the simulation very long and the errors rapidly increase with time. See Figure 3 for typical errors of about 20% on the validation set with clear signs of overfitting. https://arxiv.org/pdf/1910.07291.pdf

Paper and discussion history. A mention is not an endorsement.

Newton vs the machine: solving the chaotic three-body problem using deep neural networks — mentioned by EndXA

Sat, 26 Oct 2019 15:08:32 GMT

The original paper can be found here: https://arxiv.org/abs/1910.07291 Abstract: > Since its formulation by Sir Isaac Newton, the problem of solving the equations of motion for three bodies under their own gravitational force has remained practically unsolved. Currently, the solution for a given initialization can only be found by performing laborious iterative calculations that have unpredictable and potentially infinite computational cost, due to the system's chaotic nature. We show that an ensemble of solutions obtained using an arbitrarily precise numerical integrator can be used to train…

Paper and discussion history. A mention is not an endorsement.