Research map / Time series / Causal and generative time series
Detecting Anomalous Event Sequences with Temporal Point Processes
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 log-likelihood with the 3S statistic for anomaly scoring | RNN neural TPP · Marked sequences across Logs and simulated datasets | ROC AUC | Helped |
| Use 3S rather than KS arrival or chi-squared | TPP goodness-of-fit testing · SPP alternatives | ROC AUC | Helped |
| Use 3S rather than KS inter-event | TPP goodness-of-fit testing · SelfCorrecting SPP alternative | ROC AUC | Mixed |
| Use 3S rather than KS inter-event | RNN neural TPP anomaly detection · Server-Stop and Server-Overload | ROC AUC | Helped |
| Use 3S rather than log-likelihood | RNN neural TPP anomaly detection · Server-Stop and Server-Overload | ROC AUC | Helped |
| Use 3S rather than KS arrival or chi-squared | RNN neural TPP anomaly detection · Real-world Logs and STEAD | ROC AUC | Mixed |
| Use 3S rather than the strongest reported baseline | RNN neural TPP anomaly detection · Logs — Packet corruption (1%) | ROC AUC: 75.9 $\pm$ 0.1 -> 95.5 $\pm$ 0.3 | Helped |
| Use 3S rather than the strongest reported baseline | RNN neural TPP anomaly detection · Logs — Packet corruption (10%) | ROC AUC: 99.0 $\pm$ 0.0 -> 99.4 $\pm$ 0.1 | Helped |
| Use 3S rather than the strongest reported baseline | RNN neural TPP anomaly detection · Logs — Packet duplication (1%) | ROC AUC: 88.1 $\pm$ 0.1 -> 90.9 $\pm$ 0.3 | Helped |
| Use 3S rather than the strongest reported baseline | RNN neural TPP anomaly detection · Logs — Packet delay (frontend) | ROC AUC: 99.3 $\pm$ 0.1 -> 97.6 $\pm$ 0.1 | Hurt |
| Use 3S rather than the strongest reported baseline | RNN neural TPP anomaly detection · Logs — Packet delay (all services) | ROC AUC: 99.8 $\pm$ 0.0 -> 99.6 $\pm$ 0.1 | No clear effect |
| Use 3S rather than the strongest reported baseline | RNN neural TPP anomaly detection · STEAD — Anchorage, AK | ROC AUC: 88.0 $\pm$ 0.1 -> 88.3 $\pm$ 0.6 | No clear effect |
Closest papers by meaning
- EasyTPP: Towards Open Benchmarking Temporal Point Processes (ICLR 2024 · similarity 0.85)
- Conformal Anomaly Detection in Event Sequences (ICML 2025 · similarity 0.85)
- The Elephant in the Room: Towards A Reliable Time-Series Anomaly Detection Benchmark (NeurIPS 2024 · similarity 0.84)
- Supervision Recovery for Time Series Anomaly Detection via Counterfactual Pairing (NeurIPS 2026 · similarity 0.83)
- Residual TPP: A Unified Lightweight Approach for Event Stream Data Analysis (ICML 2025 · similarity 0.83)
- Detect, Explain, Interpret: An End-to-End Benchmark for Time Series Anomaly Detection, Explainability and Interpretability. (NeurIPS 2026 · 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.