[1906.01563] Hamiltonian Neural Networksopen searchopen navigation menucontact arXivarXiv Twitter

Even though neural networks enjoy widespread use, they still struggle to learn the basic laws of physics. How might we endow them with better inductive biases? In this paper, we draw inspiration from Hamiltonian mechanics to train models that learn and respect exact conservation laws in an unsupervised manner. We evaluate our models on problems where conservation of energy is important, including the two-body problem and pixel observations of a pendulum. Our model trains faster and generalizes better than a regular neural network. An interesting side effect is that our model is perfectly reversible in time.

2 mentions: @jinbeizame007
Date: 2020/05/11 11:21

Referring Tweets

@jinbeizame007 今日は、解析力学を少し勉強したからHamiltonian Neural Networksの論文に挑戦してみてるんだけど、思ってたよりずっと分かりやすくて理解できる! Hamiltonian NNを考えた人天才すぎませんか? t.co/jXr8QJxKzM

Related Entries

Read more はじめての自然言語処理 spaCy/GiNZA を用いた自然言語処理 | オブジェクトの広場
1 users, 0 mentions 2020/04/06 08:21
Read more International Journal of Data Science and Analytics, Volume 9, Issue 4 - Springer
1 users, 1 mentions 2020/06/26 18:52
Read more NOB DATA株式会社 | 福岡 データサイエンティスト
1 users, 2 mentions 2020/08/02 11:21
Read more GANはただの画家ではない!鑑定士への転身を果たすAnoGAN | AI-SCHOLAR | AI:(人工知能)論文・技術情報メディア
1 users, 0 mentions 2020/08/20 11:51
Read more DockerでTensorflow2のGPU環境を構築する。(Ubuntu Server) – アルカディア・システムズ株式会社 AIラボ
1 users, 1 mentions 2020/10/21 17:22