Research map / Time series / Causal and generative time series
Hierarchical ODE: Learning Continuous-Time Physical Prototypes for Early Link Failure Detection
Jiaen Lv, Leran Qi, Shaowei Wang
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 |
|---|---|---|---|
| replace discrete GRU encoder with neural ODE encoder (ablation) | hierarchical prototype network · Early link failure detection, Scenario 3 | FRR: 0.29 -> 0.00 | Improved |
| replace LSTM encoder with neural ODE encoder (ablation) | hierarchical prototype network · Early link failure detection, Scenario 6 | FRR: 0.17 -> 0.05 | Improved |
| replace ModernTCN encoder with neural ODE encoder (ablation) | hierarchical prototype network · Early link failure detection, Scenario 6 | FRR: 0.94 -> 0.05 | Improved |
| replace LSTM encoder with neural ODE encoder (ablation) | hierarchical prototype network · Early link failure detection, Scenario 2 | FAR: 0.36 -> 0.04 | No clear effect |
| use neural ODE instead of discrete encoders (ablation) | hierarchical prototype network · Early link failure detection, Scenario 1 | FAR: 0.04 vs 0.04 (GRU, ResNet) | No clear effect |
| use neural ODE instead of ResNet/ModernTCN (ablation) | hierarchical prototype network · reconstruction MSE on complete sequences | MSE | Hurt |
| use neural ODE instead of recurrent baselines (ablation) | hierarchical prototype network · reconstruction MSE on complete sequences | MSE | Improved |
| use neural ODE instead of discrete encoders under missing initial segments (ablation) | hierarchical prototype network · delayed-observation protocol, MSE | MSE | Improved |
| use neural ODE instead of ModernTCN under missing data (ablation) | hierarchical prototype network · delayed-observation protocol, FAR and FRR | FAR/FRR | Improved |
| use neural ODE latent space instead of GRU latent for hierarchical clustering (ablation) | hierarchical clustering · Scenario 1 dendrogram | prototype fragmentation: GRU: 6 clusters -> ODE: 4 clusters | Improved |
Closest papers by meaning
- Neural Context Flows for Meta-Learning of Dynamical Systems (ICLR 2025 · similarity 0.84)
- Signed Graph Neural Ordinary Differential Equation for Modeling Continuous-Time Dynamics (AAAI 2024 · similarity 0.84)
- Neural Stochastic Flows: Solver-Free Modelling and Inference for SDE Solutions (NeurIPS 2025 · similarity 0.84)
- Extracting Governing Equations from Latent Dynamics via Multi-View Contrastive Learning (NeurIPS 2026 · similarity 0.84)
- Semi-supervised Learning of Dynamical Systems with Neural Ordinary Differential Equations: A Teacher-Student Model Approach (AAAI 2024 · similarity 0.84)
- Disentangling Continuous-Time Latent Dynamics: Identifiability of Latent SDEs via Diffusion Shifts (NeurIPS 2026 · similarity 0.83)
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.
