Research map / Probabilistic and scientific modeling / Dynamical systems
Meta-iLaD: Identifiable Latent Dynamics via Meta-Learning of Dynamics Environments
Yubo Ye, Sweekar Piya, Xiajun Jiang, Linwei Wang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| infer c from few-shot context samples instead of the query sample (ablation) | Meta-iLaD · Pendulum OOD | MCC(c): 0.88 -> 0.96 | Improved |
| infer c from few-shot context samples instead of the query sample (ablation) | Meta-iLaD · Pendulum prediction | MSE(x): 5.06(0.50)e-1 -> 3.49(0.40)e-1 | Improved |
| use context samples instead of predefined environment index (ablation) | Meta-iLaD · Pendulum prediction | MSE(x): 3.45(0.22)e-1 vs 3.49(0.40)e-1 | No clear effect |
| use context samples instead of predefined labels (enables OOD) (ablation) | Meta-iLaD · Pendulum OOD | MSE(x): predefined index: not applicable; context 6.16(0.08)e-1 | Improved |
| add TC regularization (ablation) | Meta-iLaD · Pendulum prediction | MCC(c): 0.88 -> 0.98 | Improved |
| add warm-up training (L_flow) (ablation) | Meta-iLaD · Pendulum prediction | MCC(z): 0.26 -> 1.00 | Improved |
| add warm-up training (L_flow) (ablation) | Meta-iLaD · Pendulum prediction | MSE(x): 3.58(0.27)e0 -> 4.82(0.22)e-1 | Improved |
| use identifiable latent environment c with context samples vs IDF process noise | Meta-iLaD · Pendulum prediction | MCC(c): 0.02(0.01) -> 0.98(0.01) | Improved |
| use Meta-iLaD instead of IDF on OOD subject | Meta-iLaD · MoCAP OOD prediction | MSE: 1.18(0.06)e2 -> 7.87(1.25)e1 | Improved |
| identifiability of c and F matters for prediction but not reconstruction | Meta-iLaD vs IDF · Pendulum | MSE(x): reconstruction 3.78(0.34)e-1 vs 3.58(0.27)e-1; prediction 1.26(0.47)e1 vs 3.56(0.35)e-1 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- ODE2VAE — ODE2VAE: Deep generative second order ODEs with Bayesian neural networks (NeurIPS 2019) · better on 6 of 6 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Disentangling Continuous-Time Latent Dynamics: Identifiability of Latent SDEs via Diffusion Shifts (NeurIPS 2026 · similarity 0.89)
- Identifying latent state transitions in non-linear dynamical systems (ICLR 2025 · similarity 0.89)
- Learning Hybrid Dynamics Models with Simulator-Informed Latent States (AAAI 2024 · similarity 0.87)
- On the Identifiability of Switching Dynamical Systems (ICML 2024 · similarity 0.87)
- On the Identification of Temporal Causal Representation with Instantaneous Dependence (ICLR 2025 · similarity 0.87)
- Towards Identifiability of Hierarchical Temporal Causal Representation Learning (NeurIPS 2025 · similarity 0.87)
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.
