Research map / Time series / Long-term forecasting
Neural Lad: A Neural Latent Dynamics Framework for Times Series Modeling
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 attention-based signal-change transformation g_v to the h_w-only model (ablation) | Neural Lad · Synthetic series horizon 12 | MAE: 1.69 -> 1.27 | Helped |
| add attention-based g_v to the h_w-only model (ablation) | Neural Lad · Synthetic series horizon 96 | MAE: 2.28 -> 1.77 | Helped |
| add attention-based g_v to the h_w-only model (ablation) | Neural Lad · Synthetic series horizon 48 | MAE: 1.61 -> 1.47 | Helped |
| add seasonality-trend h_w(t) to the g_v-only model (ablation) | Neural Lad · Synthetic series horizon 12 | MAE: 1.44 -> 1.27 | Helped |
| add seasonality-trend h_w(t) to the g_v-only model (ablation) | Neural Lad · Synthetic series horizon 48 | MAE: 1.48 -> 1.47 | No clear effect |
| add seasonality-trend h_w(t) to the g_v-only model (ablation) | Neural Lad · Synthetic series horizon 96 | MAE: 2.07 -> 1.77 | Helped |
| replace STG-NCDE with Neural Lad | STG-NCDE · Synthetic series horizon 12 | MAE: 2.31 -> 1.27 | Helped |
| replace STG-NCDE with Neural Lad | STG-NCDE · Synthetic series horizon 96 | MAE: 3.37 -> 1.77 | Helped |
| replace Neural CDE with Neural Lad (with intensity) | Neural CDE · PhysioNet sepsis | test AUC: 0.880 -> 0.897 | Helped |
| replace Neural CDE with Neural Lad (no intensity) | Neural CDE · PhysioNet sepsis | test AUC: 0.776 -> 0.812 | Mixed |
| replace Neural CDE with Neural Lad (memory cost, with intensity) | Neural CDE · PhysioNet sepsis | memory usage (MB): 244 -> 257 | Hurt |
| replace TimesNet with Neural Lad | TimesNet · ETTh2 univariate avg | MSE: 0.414 -> 0.368 | Helped |
Closest papers by meaning
- Neural Controlled Differential Equations for Irregular Time Series (NeurIPS 2020 · similarity 0.89)
- Sequential Latent Variable Models for Few-Shot High-Dimensional Time-Series Forecasting (ICLR 2023 · similarity 0.89)
- Stable Neural Stochastic Differential Equations in Analyzing Irregular Time Series Data (ICLR 2024 · similarity 0.89)
- ARM: Refining Multivariate Forecasting with Adaptive Temporal-Contextual Learning (ICLR 2024 · similarity 0.87)
- STEM-LTS: Integrating Semantic-Temporal Dynamics in LLM-driven Time Series Analysis (AAAI 2025 · similarity 0.87)
- Neural MJD: Neural Non-Stationary Merton Jump Diffusion for Time Series Prediction (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.