Research map / Probabilistic and scientific modeling / Molecules and proteins
Learning Latent Process from High-Dimensional Event Sequences via Efficient Sampling
Qitian Wu, Zixuan Zhang, Xiaofeng Gao, Junchi Yan, Guihai 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 |
|---|---|---|---|
| replace multi-head attention with RNN (ablation) | LANTERN · Syn-Large network reconstruction (K3) | F1: 0.6253 -> 0.6220 | Hurt |
| replace multi-head attention with RNN (ablation) | LANTERN · Weibo network reconstruction (K2) | F1: 0.5663 -> 0.5631 | Hurt |
| remove discriminator and use heuristic reward (LANTERN-PR) (ablation) | LANTERN · Weibo network reconstruction (K2) | F1: 0.5663 -> 0.3692 | Hurt |
| remove discriminator and use heuristic reward (LANTERN-PR) (ablation) | LANTERN · Syn-Large network reconstruction (K3) | F1: 0.6253 -> 0.4980 | Hurt |
| remove discriminator and use heuristic reward (LANTERN-PR) (ablation) | LANTERN · Syn-Small network reconstruction (K3) | F1: 0.5733 -> 0.5823 | No clear effect |
| remove discriminator and use heuristic reward (LANTERN-PR) (ablation) | LANTERN · MemeTracker network reconstruction (K3) | F1: 0.4230 -> 0.4214 | Hurt |
| LANTERN vs better of NETRATE and KernelCascade | LANTERN · network reconstruction | F1: 14.9% | Improved |
| increase top-K edges from K1 to K3 | LANTERN · Syn-Small network reconstruction | F1: 0.4318 -> 0.5733 | Improved |
| increase top-K edges from K2 to K3 | LANTERN · Weibo network reconstruction | F1: 0.5663 -> 0.5611 | No clear effect |
| use RNN instead of attention at very large marker count (ablation) | LANTERN-RNN · Syn-Large/Weibo event prediction | time MSE and marker accuracy | Hurt |
| RNN vs heuristic-reward variant in small datasets (ablation) | LANTERN · small datasets event prediction | time MSE and marker accuracy | Mixed |
| scale markers from 100 to 100,000 | LANTERN vs NETRATE/KernelCascade · synthetic scalability | running time | Improved |
Closest papers by meaning
- User-Dependent Neural Sequence Models for Continuous-Time Event Data (NeurIPS 2020 · similarity 0.86)
- Learning Coupled Continuous-Time Latent Dynamics from Irregular Events (ICML 2026 · similarity 0.86)
- Interacting Diffusion Processes for Event Sequence Forecasting (ICML 2024 · similarity 0.85)
- A Variational Autoencoder for Neural Temporal Point Processes with Dynamic Latent Graphs (AAAI 2024 · similarity 0.85)
- Sequence Modeling with Spectral Mean Flows (NeurIPS 2025 · similarity 0.84)
- Neural Temporal Point Processes for Forecasting Directional Relations in Evolving Hypergraphs (AAAI 2025 · similarity 0.83)
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.
