401-4656-21L AI in the Sciences and Engineering
| Semester | Spring Semester 2024 |
| Lecturers | S. Mishra, B. Moseley |
| Periodicity | yearly recurring course |
| Language of instruction | English |
| Comment | Aimed at students in a Master's Programme in Mathematics, Engineering and Physics. |
| Abstract | AI is having a profound impact on science by accelerating discoveries across physics, chemistry, biology, and engineering. This course aims to present a highly topical selection of AI applications across these fields. Emphasis will be placed on using AI, particularly deep learning, to understand systems modelled by PDEs, and key scientific machine learning concepts and themes will be discussed. | |||||||||||||||||||||||||||||||||
| Learning objective | Learning objectives: - Aware of advanced applications of AI in the sciences and engineering - Familiar with the design, implementation, and theory of these algorithms - Understand the pros/cons of using AI and deep learning for science - Understand key scientific machine learning concepts and themes | |||||||||||||||||||||||||||||||||
| Content | A selection of the following topics will be presented in the lectures: 1. Key scientific tasks common to many scientific domains, such as simulation, inverse problems, equation discovery, design, and control problems, and issues with traditional methods for solving them 2. Physics-informed neural networks for solving forward, inverse and equation discovery problems related to PDEs 3. Neural operators, including Fourier neural operators and DeepONets, for learning efficient surrogate models, and their theoretical foundations 4. Differentiable scientific algorithms, neural differential equations, and the benefits of hybrid workflows 5. AI for symbolic regression and equation discovery 6. Applications of graph neural networks in science 7. Guest lectures on AI for chemistry and biology 8. Large language models and other Foundation models for scientific discovery Applications using these techniques will be illustrated across fluid dynamics, wave physics, medical physics, molecular design, and computational biology. Several examples where AI algorithms outperform traditional scientific workflows will be shown. | |||||||||||||||||||||||||||||||||
| Lecture notes | Lecture slides, recordings, and tutorials will be available on Moodle. | |||||||||||||||||||||||||||||||||
| Literature | All the material in the course is based on research articles written in last 1-3 years. The relevant references will be provided. | |||||||||||||||||||||||||||||||||
| Prerequisites / Notice | - An understanding of basic ML concepts including supervised learning, overfitting/underfitting, optimisation, and neural networks is required; you should understand the main concepts in the ETH 252-0220-00L Introduction to Machine Learning course (but note, completing this course is not a formal requirement) - Familiar with PDEs and numerical methods for solving them - Basic competence in Python and some practical familiarity with deep learning frameworks (e.g. PyTorch, TensorFlow, or Keras) | |||||||||||||||||||||||||||||||||
Competencies |
|