https://hackerpaper.kaipfister.de/p/1909.12790/Collected HN comments and submissions linking this paper.enMon, 02 Nov 2026 11:16:22 GMThttps://news.ycombinator.com/item?id=21365587urn:hackerpapers:mention:1909.12790:21365587Sat, 26 Oct 2019 19:14:20 GMT<p>Agreed, this paper did not show any meaningful generalization and is sorely short on baselines. If you wanted to train a neural network model that would generalize, you would need build in physical constraints into the model architecture. For example, dynamics are governed by a Hamiltonian based on pairwise interactions, which should be integrated as an ODE. A nice recent example of this from folks at DeepMind is this paper “Hamiltonian Graph Networks with ODE Integrators”: https://arxiv.org/abs/1909.12790. That said, if you go down this road too far you may find that you are just learning an…</p><p><a href="https://hackerpaper.kaipfister.de/p/1909.12790/">Paper and discussion history</a>. A mention is not an endorsement.</p>cs.LG