Research map / Probabilistic and scientific modeling / Causal discovery
Causal Inference for Event Pairs in Multivariate Point Processes
Tian Gao, Dharmashankar Subramanian, Debarun Bhattacharjya, Xiao Shou, Nicholas Mattei, Kristin Bennett
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 stable weights (IP-Stable) instead of conditional intensity score CI | IPTW ATE estimator · Diabetes test, K=10 | Hits@K: 1 -> 3 | Improved |
| IP-Stable vs CIM baseline | IPTW ATE estimator · Diabetes test, K=10 | Hits@K: 4 -> 3 | Hurt |
| IP-Stable vs CIM baseline | IPTW ATE estimator · Diabetes test, K=15 | Hits@K: 5 -> 6 | Hurt |
| IP-Stable vs CI | IPTW ATE estimator · Diabetes test, K=20 | Hits@K: 3 -> 7 | Improved |
| use non-stable weights (IP-NS) instead of stable weights (ablation) | IPTW · Synthetic PGEM datasets | RMSE | Hurt |
| apply stable weights (IP-Stable) instead of CI scores | IPTW · Synthetic PGEM datasets | RMSE: up to 6 times reduction in error | Improved |
| IP-Stable instead of CI scores | IPTW · Synthetic Hawkes datasets | RMSE: 3 to 10 times better accuracy | Improved |
| no weighting (IP-NW) instead of CI scores (ablation) | IPTW · Synthetic PGEM, small window | RMSE | Hurt |
| no weighting (IP-NW) instead of CI scores | IPTW · Synthetic Hawkes | RMSE | Improved |
| CIP-NS non-stable weights vs CIP-Stable (ablation) | CIP · PGEM-1, future window 15 | ATE RMSE: 3.11±2.42 -> 122.6±60.6 | Hurt |
| CIP-NS non-stable weights vs CIP-Stable (ablation) | CIP · PGEM-2, future window 15 | ATE RMSE: 2.87±0.99 -> 538.91±203.36 | Hurt |
| CIP-NW instead of CCI | CIP · PGEM-1, future window 15 | ATE RMSE: 4.51±0.94 -> 3.20±2.65 | Improved |
Closest papers by meaning
- A Practical Approach to Causal Inference over Time (AAAI 2025 · similarity 0.87)
- CausalRivers - Scaling up benchmarking of causal discovery for real-world time-series (ICLR 2025 · similarity 0.87)
- On the Recoverability of Causal Relations from Temporally Aggregated I.I.D. Data (ICML 2024 · similarity 0.86)
- Discovering Mixtures of Structural Causal Models from Time Series Data (ICML 2024 · similarity 0.86)
- A Dynamical View of the Question of Why (ICLR 2024 · similarity 0.86)
- Causal Identification for Complex Functional Longitudinal Studies (ICLR 2025 · similarity 0.86)
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.
