Research map / Probabilistic and scientific modeling / Dynamical systems
Symplectic Recurrent Neural Networks
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add recurrent multi-step training to L-L H-NET (ablation) | H-NET · Spring-chain noisy | position L2 error mean: 3.36 -> 2.37 | Helped |
| add recurrent multi-step training to E-E H-NET (ablation) | H-NET · Spring-chain noisy | position L2 error mean: 7.24 -> 7.58 | Hurt |
| use leapfrog for both training and testing instead of Euler (ablation) | recurrent H-NET · Spring-chain noisy | position L2 error mean: 7.58 -> 2.37 | Helped |
| add initial state optimization (ablation) | SRNN · Spring-chain noisy | position L2 error mean: 2.37 -> 1.45 | Helped |
| add recurrent multi-step training to L-L H-NET (ablation) | H-NET · Three-body | prediction error mean: 0.35 -> 0.26 | Helped |
| use leapfrog for both training and testing instead of Euler (ablation) | recurrent H-NET · Three-body | prediction error mean: 0.79 -> 0.26 | Helped |
| learn the dynamics from data rather than simulate the true equations at the same step size | SRNN · Three-body | prediction error mean: 0.47 -> 0.26 | Helped |
| train and test with the same integrator and time-step (ablation) | Hamiltonian neural networks · Spring-chain noisy | position L2 error mean: 5.26 -> 2.37 | Mixed |
Closest papers by meaning
- Sparse Symplectically Integrated Neural Networks (NeurIPS 2020 · similarity 0.90)
- Hyperverlet: A Symplectic Hypersolver for Hamiltonian Systems (AAAI 2022 · similarity 0.89)
- Stiffness-aware neural network for learning Hamiltonian systems (ICLR 2022 · similarity 0.88)
- Mechanistic Interpretability of RNNs emulating Hidden Markov Models (NeurIPS 2025 · similarity 0.85)
- Learning long range dependencies through time reversal symmetry breaking (NeurIPS 2025 · similarity 0.85)
- Rethinking Human Motion Prediction with Symplectic Integral (CVPR 2024 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.