Research map / Label-efficient and robust learning / Recognition and long-tail
A Distributed Multi-Sensor Machine Learning Approach to Earthquake Early Warning
Kevin Fauvel, Daniel Balouek-Thomert, Diego Melgar, Pedro Silva, Anthony Simonet, Gabriel Antoniu, Alexandru Costan, Véronique Masson, Manish Parashar, Ivan Rodero, Alexandre Termier
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 |
|---|---|---|---|
| combine GPS and seismometer sensor predictions with 1-NN stacking instead of seismometer-only argmax | DMSEEW · Earthquake early warning (3-fold CV, 269 events) | accuracy %: 73.2 -> 76.9 | Improved |
| stacked 1-NN combination instead of rule of relative strength | DMSEEW · Earthquake early warning (3-fold CV, 269 events) large earthquakes | F1 %: 72.7 -> 100.0 | Improved |
| stacked 1-NN combination instead of seismometer-only baseline | DMSEEW · Earthquake early warning (3-fold CV, 269 events) large earthquakes | F1 %: 55.2 -> 100.0 | Improved |
| stacked 1-NN combination instead of rule of relative strength | DMSEEW · Earthquake early warning (3-fold CV, 269 events) medium earthquakes | F1 %: 46.0 -> 51.6 | Improved |
| stacked 1-NN combination instead of seismometer-only baseline | DMSEEW · Earthquake early warning (3-fold CV, 269 events) medium earthquakes | F1 %: 45.0 -> 51.6 | Improved |
| stacked combination instead of rule of relative strength | DMSEEW · Earthquake early warning (3-fold CV, 269 events) normal activity | F1 %: 84.4 -> 83.8 | No clear effect |
| use WEASEL+MUSE instead of MLSTM-FCN as sensor-level classifier (ablation) | DMSEEW step 1 · Seismometer sensor-level | accuracy %: 54.6 -> 63.6 | Improved |
| use WEASEL+MUSE instead of MLSTM-FCN as sensor-level classifier (ablation) | DMSEEW step 1 · GPS sensor-level | accuracy %: 98.9 -> 99.5 | Improved |
| use WEASEL+MUSE instead of DTWD/DTWI (ablation) | DMSEEW step 1 · Seismometer sensor-level | accuracy %: 35.3 -> 63.6 | Improved |
| use 1-NN instead of Elastic Net, SVM, RF or XGBoost at central level (ablation) | DMSEEW step 2 · Earthquake early warning (3-fold CV, 269 events) | accuracy % | Improved |
Closest papers by meaning
- Broadband Ground Motion Synthesis by Diffusion Model with Minimal Condition (ICML 2025 · similarity 0.79)
- Early Detection of Extreme Storm Tide Events Using Multimodal Data Processing (AAAI 2024 · similarity 0.78)
- A Scalable Multi-Task Model for Virtual Sensors (NeurIPS 2026 · similarity 0.78)
- AI-Based Energy Transportation Safety: Pipeline Radial Threat Estimation Using Intelligent Sensing System (AAAI 2024 · similarity 0.77)
- Accident Anticipation via Temporal Occurrence Prediction (NeurIPS 2025 · similarity 0.77)
- MONITRS: Multimodal Observations of Natural Incidents Through Remote Sensing (NeurIPS 2025 · similarity 0.77)
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.
