Skip to content

🔄 Meta-Learning Regime Shift

This simulation extends the nested-learning two-state model by adding regime shifts and a meta-learner that adapts the learning rate in response to prediction error.

🧠 Idea

A two-state Markov system changes its transition probability abruptly every N steps (regime shift). Two agents learn the system dynamics side by side:

Agent Learning Rate
Fixed-LR Constant η – good baseline
Meta-Learning η is adjusted dynamically: high surprise → raise η, low surprise → lower η

The meta-learner demonstrates Adaptive Capacity (A) from the System Intelligence Index: the ability to change one's own learning behaviour when conditions change.

đŸ–ŧ Visualisation

A 3-panel matplotlib figure:

  1. Prediction error over time – both agents, with regime-shift markers
  2. Learning rate Ρ of the meta-learner (log scale)
  3. Learned vs. true transition probability – convergence tracking

â–ļ Run

cd simulation-models/meta-learning-regime-shift
python3 meta_regime_shift.py