ME 539 — Introduction to Scientific Machine Learning

This repository contains the lecture book, Jupyter notebooks, and supporting materials for Purdue University's ME 539, Introduction to Scientific Machine Learning. The current offering is taught by Prof. Ilias Bilionis in Fall 2026.

Fall 2026 offering

• Campus students meet Tuesdays and Thursdays, 10:30–11:15 a.m. ET, in ME 2061.
• Online/EPE and asynchronous students have no required synchronous class meetings. They complete the same course modules and follow the published homework and exam deadlines.
• The course uses 28 numbered lecture modules, ten homework assignments, and three noncomprehensive unit exams.
• Registered Purdue students access lecture videos, announcements, assignment links, and the official course schedule through Brightspace. Homework is submitted through Gradescope.
• Google Colab is the recommended environment for course activities and homework. The repository can also be run locally using the locked uv environment described below.

Course policies, deadlines, examination logistics, and accommodations are governed by the current syllabus and Brightspace site.

The material is distributed under the GNU General Public License v3. If you reuse or modify it, follow the terms in LICENSE and cite this repository. Prof. Bilionis welcomes a note at ibilion@purdue.edu when the material is reused.

Lecture Book

The published lecture book is available here. It contains the reading activities, hands-on notebooks, and homework pages used by the course. Pre-recorded videos for the numbered lecture modules are linked through the Purdue course site.

Local development

All Python environments, dependencies, and command-line tools for this repository are managed with uv:

uv sync --locked

The notebooks that draw directed graphs also require the Graphviz dot system executable. On macOS it can be installed with brew install graphviz; this is a system dependency rather than a Python package.

To audit notebook execution without changing their stored outputs, run:

uv run --locked python tools/execute_notebooks.py

Pass --in-place only when the refreshed outputs should be saved. Build the classic Jupyter Book with ./make.sh; maintainers can publish the validated build with ./make.sh --publish.

Older Course Versions

• Fall 2020 version
• Fall 2021 version
• Summer 2022 version
• Fall 2023 version

This course evolved from the ME 597 "Data Analytics for Scientists and Engineers," taught two times by Prof. Bilionis, and the ME 597/MA 598 "Introduction to Uncertainty Quantification," taught three times by Prof. Bilionis (the first time, Spring 2016 it was co-taught with Prof. Guang Lin). If you are interested in accessing the old versions of the course, they can be found here.

Note that there is also a 1-credit undergraduate version of the course under ME 297, "Introduction to Data Science for Mechanical Engineers." This version can be found here.