University of Pennsylvania - Spring 2025
Instructor: Paris Perdikaris
This repository contains all course materials for MEAM4600: AI for Science and Engineering: From Data to Discovery, an undergraduate course introducing machine learning techniques tailored for science and engineering applications. The repository is updated weekly as we progress through the semester.
- Lecture Slides: Weekly presentation materials covering foundational concepts and applications
- Jupyter Notebooks: Interactive tutorials and hands-on exercises
- Code Examples: Python implementations of key algorithms and methods
- Homework Assignments: 6 problem sets throughout the semester
- Assignment Solutions: Released after submission deadlines
- Project Guidelines: Templates and requirements for the final project
- Primary Textbook: Murphy, K. P. (2022). Probabilistic Machine Learning: An Introduction. MIT Press
- Supplementary Papers: Research articles and case studies
- Additional Resources: Tutorials and reference materials
Embark on an exciting journey into the world of AI-driven scientific discovery and engineering innovation. This course introduces cutting-edge machine learning techniques tailored for science and engineering applications. From enhancing material properties to revolutionizing drug discovery, from optimizing renewable energy systems to advancing robotics, AI is reshaping how we approach complex problems across disciplines.
Students will gain practical expertise in:
- Machine Learning Fundamentals: Regression, classification, and statistical estimation
- Deep Learning: MLPs, CNNs, RNNs, GNNs, and Transformers for engineering applications
- Physics-Informed ML: Neural networks and operators that incorporate physical laws
- Probabilistic Methods: Variational inference, Monte Carlo sampling, Bayesian optimization
- Generative Models: VAEs and modern generative approaches
- Uncertainty Quantification: Gaussian processes and active learning
Through hands-on projects and case studies, you'll apply AI to:
- Materials discovery and property prediction
- Computational fluid dynamics and engineering design
- Drug discovery and biological systems
- Renewable energy optimization
- Robotics and autonomous systems
- Mathematics: MATH 240 (Linear Algebra)
- Programming: ESE 2030 or ENGR 1050 (Python programming)
- No prior machine learning experience required
This course is cross-listed across multiple departments:
- MEAM 4600 - Mechanical Engineering and Applied Mechanics
- EE 4600 - Electrical Engineering
- CBE 4600 - Chemical and Biomolecular Engineering
- MSE 4600 - Materials Science and Engineering
- Python: Anaconda distribution (recommended)
- ML Libraries: NumPy, SciPy, scikit-learn, JAX, PyTorch
- Development Environment: Jupyter Notebook or JupyterLab
- Cloud Computing: Google Colab for GPU access
| Week | Topic |
|---|---|
| 1 | Introduction to AI in Science and Engineering |
| 2 | Fundamentals of Machine Learning for Scientific Computing |
| 3 | Primer on Probability and Statistical Estimation |
| 4 | Linear & Logistic Regression |
| 5 | Variational Inference |
| 6 | Monte Carlo Sampling |
| 7 | Deep Learning Architectures: MLPs, CNNs |
| 8 | Deep Learning Architectures: RNNs, GNNs, Transformers |
| 9 | Physics-Informed Neural Networks & Neural Operators |
| 10 | Kernel Methods & Gaussian Process Regression |
| 11 | Bayesian Optimization and Active Learning |
| 12 | Unsupervised Learning: PCA and VAEs |
| 13 | Generative Models |
| 14 | Final Project Presentations |
- Homework Assignments (40%): 6 problem sets
- Midterm Exam (20%): In-class examination
- Final Project (40%): Team-based applied ML project
MEAM4600/
├── lectures/ # Weekly lecture slides
├── notebooks/ # Jupyter tutorials and exercises
├── assignments/ # Homework problems and solutions
├── project/ # Final project guidelines and templates
└── resources/ # Additional materials and references
-
Clone the repository:
git clone https://github.com/[username]/MEAM4600.git cd MEAM4600 -
Set up Python environment:
conda create -n meam4600 python=3.10 conda activate meam4600 pip install -r requirements.txt
-
Launch Jupyter:
jupyter notebook
-
Test your setup with the Week 1 notebook to ensure all dependencies are installed correctly.
This repository is updated weekly with:
- New lecture materials and slides
- Jupyter notebooks with coding exercises
- Homework assignments and starter code
- Additional resources and readings
- Solutions to previous assignments (after deadline)
Upon completion of this course, students will be able to:
- Apply advanced machine learning techniques to analyze and model complex systems in engineering and scientific domains
- Develop AI-driven solutions for real-world problems in their specific fields of study
- Critically evaluate the potential and limitations of AI methods in scientific discovery and engineering design
- Implement and fine-tune state-of-the-art machine learning models using modern software frameworks
- Communicate the results and implications of AI-driven analyses effectively to both technical and non-technical audiences
- Instructor: Paris Perdikaris
- Office Hours: Tuesdays & Thursdays, 3-4pm, Rm 521, Amy Gutmann Hall (or by appointment)
- Course Website: https://www.seas.upenn.edu/~meam4600/
- Canvas: https://canvas.upenn.edu/courses/1912998
- Discussion Forum: https://edstem.org/us/courses/93496/
Students are expected to follow Penn's Code of Academic Integrity. Collaboration on homework is encouraged, but each student must write and submit their own solutions. All sources must be properly cited.
Students are encouraged to:
- Report issues or typos in course materials
- Suggest improvements to code examples
- Share interesting applications or extensions of course concepts
If you need accommodations for a disability, please contact Student Disabilities Services and inform the instructor as early as possible in the semester.
Course materials are provided for educational use by enrolled students. Please respect intellectual property rights and university policies regarding academic integrity.
This repository supports MEAM4600: AI for Science and Engineering at the University of Pennsylvania. Materials are updated regularly throughout the semester.
Course Catalog: This course satisfies Math requirements and is open to undergraduates with appropriate prerequisites.