Advanced RL Code (Neural Data Science)

Notebooks and code for contextual bandit simulation and RPE decoding with LaseNet.

Run locally

1. Clone the repo (without the large .venv if you added it to .gitignore).
2. Create and activate a virtual environment:

   python -m venv .venv
   source .venv/bin/activate   # Windows: .venv\Scripts\activate

3. Install dependencies:

   pip install -r requirements.txt

4. Start Jupyter and open the notebooks:

   jupyter notebook

Contents

• decoding-example.ipynb – Decoding reward prediction errors with LaseNet.
• simulate-contextual-bandit.ipynb – Contextual bandit simulation and attention exercises.
• simulate-contextual-bandit-solutions.ipynb – Solutions for the bandit notebook.
• libraries/ – World, FeatureRL, Data, plotting for simulations.