Introduction to Reinforcement Learning (in HCI)

author: Thomas Langerak

This is a notebook that will be taught at the Computational Interaction Summer School 2025. The notebook is accompanied by a slide-deck, and the text in the notebook is more as reference than meant to be complete. The slides are partially based on those of Sammy Christen, at ETH Zurich.

Installation

to install uv simple run:

pip install uv

I have tested with python 3.12, so I recommend using that version: uv python install 3.12

create uv .venv

uv venv .venv --python 3.12

activate uv .venv

bash: source .venv/bin/activate

windows: .venv\Scripts\activate

install requirements

uv pip install -r requirements.txt

or manually:

uv pip install mujoco glfw numpy pillow matplotlib "gymnasium[other]" "stable-baselines3[extra]" tensorboard tqdm seaborn ipywidgets gymnasium

test

to start the notebook, run the following in the main folder:

uv run jupyter notebook intro_to_rl_answers.ipynb

this should open a folder in your browser, otherwise go to

localhost:8888

then open intro_to_rl.ipynb and run the cell with all the imports (right below motivation)

if everything imports without errors, we are all set.