How AI Is Used
LSTM neural networks learn spatio-temporal dependencies in earthquake sequences - here is what that means in practice.
The Role of AI in Seismikon
Seismikon uses Long Short-Term Memory (LSTM) neural networks - a class of recurrent neural networks - as the core prediction engine. The LSTM learns patterns in historical earthquake sequences and uses those patterns to generate forward predictions for location, magnitude, depth, and time window.
The AI is one component of a pipeline that also includes natural time preprocessing, sliding window feature extraction, and the EPS calculation. The LSTM does not run directly on raw seismic waveforms; it runs on structured features derived from the earthquake catalogue.
Why LSTM Networks for Seismicity
Earthquake sequences have long-range temporal dependencies: the likelihood of a large event may be influenced by activity patterns months or years earlier. Standard feedforward networks struggle with this because they have no memory of past inputs. LSTM networks are specifically designed to retain and selectively use information from long sequences.
- Long-term dependencies: LSTM memory cells retain information from many time steps back - critical for seismic cycle patterns.
- Spatio-temporal learning: the network learns correlations between earthquakes at different locations and times simultaneously.
- Multi-scale patterns: the sliding window input exposes the model to both short-term precursory signals and longer-term background trends.
- Sequence output: Connectionist Temporal Classification (CTC) training allows the model to produce sequences of predictions rather than single-event forecasts.
What the Model Predicts
The LSTM output is a 4-tuple: predicted latitude, longitude, focal depth (km), and magnitude. Each prediction also carries a model confidence score and the EPS value at prediction time. The time window is fixed at 48 hours from prediction issuance.
The model is trained on California earthquake sequences from 2000–2023 (Bikos 2024) and deployed on real-time USGS catalogue data.
Limitations of the AI Approach
Machine learning models learn from historical patterns. If the seismic system enters a regime genuinely unlike the training data - an entirely new fault segment activating, a rare large-magnitude event with no historical analogue - the model's predictions may be less reliable. This is a known limitation of all data-driven seismic forecasting approaches.
The AI is a tool for surfacing patterns in data; it is not a physical model of fault mechanics. See Limitations & Open Questions for the full account.