For instructors

The dictionary is an open (CC BY 4.0) course resource: 68 machine-learning terms, each defined precisely, cross-referenced, and shipped with a typeset PDF and, for most terms, a Python script that recomputes what the entry states. Everything below can be used in a course without asking; attribution is the only condition.

Build a reading list

The Seminar-Baukasten is a configurator over the published terms: pick the terms a seminar needs and it assembles the deep links. Alternatively, link terms directly — every page has a stable URL of the form https://dictionaryofml.org/terms/<term>.html — or hand out the per-term PDFs, which are the authoritative typeset form of each entry.

Exercises from the demos

Most terms carry a self-contained Python demo (NumPy/Matplotlib only, fixed seeds): it states one check per claim of the entry and prints one line per check. Each demo page offers the script, a generated notebook, and an "Open in Colab" link, so students can run and modify it without installing anything. A natural exercise form: break a check, explain why it fails. Before release, every demo is screened with ruff (correctness rules) and bandit (exec/eval/shell checks) on top of its own assertions and the project's figure-style linters: screens, not proofs of correctness.

Use the notation

The book's notation — one macro per concept, one meaning per symbol — is packaged as mldict.sty: \usepackage{mldict} in course slides or theses gives the same symbols the dictionary uses, and the List of symbols is the reference for what each one means.

Plug it into AI assistants

The full dictionary is machine-readable: terms.json (one structured file, every entry) and llms.txt (a curated index for LLM crawlers). The companion repository carries an MCP server exposing lookup, search and related-term tools, so Claude Code, VS Code Copilot and similar assistants can consult the dictionary in context — useful for enforcing consistent terminology in student projects.

Cite it

@misc{dictml,
  author = {Jung, Alexander and Olioumtsevits, Konstantina and Schnoor, Ekkehard},
  title = {Dictionary of Applied Machine Learning (course edition)},
  year = {2026},
  doi = {10.5281/zenodo.21569296},
  note = {ISBN 978-952-64-3013-3, CC BY 4.0},
  url = {https://dictionaryofml.org}
}

Each term page also carries its own "Cite this entry" BibTeX block.

Adopting it in a course?

Corrections, suggestions and adoption notes are welcome: alexjung235@gmail.com. Knowing where the dictionary is used helps keep it funded and maintained.