Research map / Probabilistic and scientific modeling / Variational and generative sampling
Wasserstein Learning of Deep Generative Point Process Models
Shuai Xiao, Mehrdad Farajtabar, Xiaojing Ye, Junchi Yan, Le Song, Hongyuan Zha
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 Wasserstein training instead of MLE with the same RNN architecture | RNN point-process generator · NN synthetic process | intensity deviation: 0.341 (3.4e-1) -> 0.144 (4.28e-2) | Improved |
| Use Wasserstein training instead of MLE with the same RNN architecture | RNN point-process generator · NN synthetic process | QQ plot slope deviation: 0.065 (1.2e-2) -> 0.012 (4.7e-3) | Improved |
| Use a Wasserstein generator rather than the strongest listed MLE estimator | Point-process modeling · SE synthetic process | QQ plot slope deviation: 0.001 (1.3e-6) -> 0.051 (1.8e-3) | Hurt |
| Use a Wasserstein generator rather than the strongest listed MLE estimator | Point-process modeling · SC synthetic process | intensity deviation: 0.006 (9.9e-5) -> 0.302 (2.2e-3) | Hurt |
| Use Wasserstein learning instead of parametric MLE estimators on mixed-process data | Point-process modeling · IP+SE+SC | intensity deviation: 0.410 (1.8e-5) -> 0.305 (6.1e-2) | Improved |
| Use Wasserstein learning instead of parametric MLE estimators on mixed-process data | Point-process modeling · IP+SC+NN | intensity deviation: 0.713 (1.7e-1) -> 0.525 (2.2e-3) | Improved |
| Use Wasserstein learning instead of parametric MLE estimators on mixed-process data | Point-process modeling · IP+SE+NN | intensity deviation: 0.678 (3.4e-1) -> 0.419 (4.2e-2) | Improved |
| Use Wasserstein learning instead of parametric MLE estimators on mixed-process data | Point-process modeling · SE+SC+NN | intensity deviation: 0.286 (3.6e-1) -> 0.200 (3.8e-2) | Improved |
| Use Wasserstein point-process learning rather than MLE baselines | Point-process modeling · MIMIC | intensity deviation: 0.150 -> 0.122 | Improved |
| Use Wasserstein point-process learning rather than MLE baselines | Point-process modeling · Meme | intensity deviation: 0.701 -> 0.351 | Improved |
| Use Wasserstein point-process learning rather than MLE baselines | Point-process modeling · MAS | intensity deviation: 1.089 -> 0.849 | Improved |
| Use Wasserstein point-process learning rather than MLE baselines | Point-process modeling · NYSE | intensity deviation: 0.347 -> 0.303 | Improved |
Closest papers by meaning
- Learning Conditional Generative Models for Temporal Point Processes (AAAI 2018 · similarity 0.89)
- UNIPoint: Universally Approximating Point Processes Intensities (AAAI 2021 · similarity 0.86)
- Wasserstein Convergence Guarantees for a General Class of Score-Based Generative Models (JMLR 2025 · similarity 0.86)
- Intensity-Free Learning of Temporal Point Processes (ICLR 2020 · similarity 0.86)
- Unlocking Point Processes through Point Set Diffusion (ICLR 2025 · similarity 0.84)
- Statistically Optimal Generative Modeling with Maximum Deviation from the Empirical Distribution (ICML 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.
