Advanced Applied Deep Learning in Physics and Engineering
Syllabus, Master's level, 1FA006
- Code
- 1FA006
- Education cycle
- Second cycle
- Main field(s) of study and in-depth level
- Physics A1F, Technology A1F
- Grading system
- Fail (U), Pass (3), Pass with credit (4), Pass with distinction (5)
- Finalised by
- The Faculty Board of Science and Technology, 29 February 2024
- Responsible department
- Department of Physics and Astronomy
General provisions
In this course, you will delve into advanced concepts in neural networks and deep learning. You will explore techniques such as Graph Neural Networks, Generative models, quantized networks, and more, along with practical skills in using tools like TensorFlow, PyTorch, and JAX. These topics will be illuminated with examples from current research in physics and technology. Upon completion of the course, you will be able to design custom neural network architectures for problems in physics and technology, handle complex datasets for training, and choose the right deep learning tools for different problems, making you ready for advanced applications in these fields.
Entry requirements
120 credits in science/engineering. Applied Deep Learning in Physics and Engineering. Proficiency in English equivalent to the Swedish upper secondary course English 6.
Learning outcomes
On completion of the course, the student should be able to:
- create and design different advanced network architectures within Keras and PyTorch for problems in physics and engineering
- read, understand and handle complex data sets and prepare them for the training of neural networks
- identify the right tool to tackle specific deep learning problems and apply the tools to the problems in the correct way
Content
Advanced deep learning architectures and concepts such as Graph Neural Networks, Transformers, Generative models (GANs and diffusion models), Normalizing Flows, uncertainty predictions, quantized networks, large language models, and text-to-image. Interpretability, ethics, and sustainability of deep learning. Practical skills in using common deep learning libraries and tools (TensorFlow/KERAS, PyTorch, JAX, scikit-learn, tensorboard, data loaders and data pipelines). All topics will be studied with examples from state-of-the-art research in physics and engineering.
Instruction
Lectures and exercise classes
Assessment
Hand-in problems