Research map / Time series / Causal and generative time series
The Arrow of Time in Multivariate Time Series
Stefan Bauer, Bernhard Schölkopf, Jonas Peters
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 |
|---|---|---|---|
| Use non-Gaussian innovations (time-direction test via multivariate VAR residual independence) versus Gaussian innovations | VARMA / VAR processes · Simulated VAR processes | direction detection | Improved |
| Model the series jointly as a multivariate VAR instead of fitting each univariate series separately (ablation) | Algorithm 1 on GDP growth data · UK, Canada, USA GDP growth | direction decision | Improved |
| Increase number of Gaussian error dimensions (from 1 non-Gaussian component up to all Gaussian) (ablation) | Algorithm 1 on 5-dim VAR · Simulated 5-dim VAR, 1000 points | direction detection | Hurt |
| Increase VAR dimension when using LiNGAM baseline | LiNGAM time-direction baseline · Simulated VAR | accuracy | Hurt |
| Apply method to NASDAQ daily returns (model misspecification) | Algorithm 1 on stock returns · NASDAQ stocks, k=3,5,10,50 | decision rate | No clear effect |
| Test on video snippets (Pickup et al. setting) | Algorithm 1 (version b) · video snippets | direction detection | No clear effect |
Closest papers by meaning
- Arrow: Accelerator for Time Series Causal Discovery with Time Weaving (ICML 2025 · similarity 0.86)
- Flow based approach for Dynamic Temporal Causal models with non-Gaussian or Heteroscedastic Noises (NeurIPS 2025 · similarity 0.85)
- GCAD: Anomaly Detection in Multivariate Time Series from the Perspective of Granger Causality (AAAI 2025 · similarity 0.84)
- A Practical Approach to Causal Inference over Time (AAAI 2025 · similarity 0.84)
- Chimera: Effectively Modeling Multivariate Time Series with 2-Dimensional State Space Models (NeurIPS 2024 · similarity 0.84)
- Rethinking Channel Dependence for Multivariate Time Series Forecasting: Learning from Leading Indicators (ICLR 2024 · similarity 0.84)
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.
