Research map / Time series / Causal and generative time series
DualTimesField: Rethinking Time Series as Continuous-Time Trends and Events
Wencheng Zhang, Long Li, Huayi Qin, Zongjuan Wu, Jing Li, Wanghu Chen
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 |
|---|---|---|---|
| add the Discrete Geometric Field (Gabor event atoms) to the trend-only Continuous Time Field (ablation) | DualTimesField · nine reconstruction benchmarks (average) | MSE: 0.121 -> 0.056 | Improved |
| add the Continuous Time Field (trend) to the events-only Discrete Geometric Field (ablation) | DualTimesField · nine reconstruction benchmarks (average) | MSE: 0.071 -> 0.056 | Improved |
| add the Continuous Time Field to events-only model on Exchange (ablation) | DualTimesField · Exchange reconstruction | MSE: 0.113 -> 0.046 | Improved |
| add the Continuous Time Field to events-only model on ETTh2 and ILI (ablation) | DualTimesField · ETTh2 / ILI reconstruction | MSE: 0.039 -> 0.040 (ETTh2); 0.050 -> 0.051 (ILI) | No clear effect |
| use dual-field trend-event decomposition instead of a single-field INR | DualTimesField · ETTm2 reconstruction vs TimesNet | MSE: 0.2424 -> 0.0075 | Improved |
| use DualTimesField on trend-dominated climate data | DualTimesField · USHCN interpolation | MSE (x10^-3): LS4 0.06 vs ours 0.17 | Hurt |
| use DualTimesField for interpolation on event-rich irregular data | DualTimesField · PhysioNet interpolation | MSE (x10^-3): 0.63 -> 0.30 (vs LS4) | Improved |
| use DualTimesField reconstruction vs iTransformer on Exchange MAE | DualTimesField · Exchange reconstruction | MAE: 0.1325 -> 0.1415 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- TimesNet — TimesNet: Temporal 2D-Variation Modeling for General Time Series Analysis (ICLR 2023) · better on 8 of 8 shared results
- iTransformer — iTransformer: Inverted Transformers Are Effective for Time Series Forecasting (ICLR 2024) · better on 2 of 3 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
- DualDynamics: Synergizing Implicit and Explicit Methods for Robust Irregular Time Series Analysis (AAAI 2025 · similarity 0.86)
- Addressing Spatial-Temporal Heterogeneity: General Mixed Time Series Analysis via Latent Continuity Recovery and Alignment (NeurIPS 2024 · similarity 0.85)
- Dynamical modeling of nonlinear latent factors in multiscale neural activity with real-time inference (NeurIPS 2025 · similarity 0.85)
- TimeMixer++: A General Time Series Pattern Machine for Universal Predictive Analysis (ICLR 2025 · similarity 0.85)
- TimePerceiver: An Encoder-Decoder Framework for Generalized Time-Series Forecasting (NeurIPS 2025 · similarity 0.85)
- TimeDART: A Diffusion Autoregressive Transformer for Self-Supervised Time Series Representation (ICML 2025 · 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.
