This course introduces modern deep‑learning methods and their application to physics problems. Building on the statistical background provided by PHYS 3243 and the mathematical and machine‑learning foundations developed in PHYS 3245, the course focuses on how neural‑network‑based models are constructed, trained, and applied. Topics include automatic differentiation, backpropagation, gradient‑based optimisation, and a range of neural‑network architectures such as fully connected and convolutional networks. The course also provides an introduction to reinforcement learning and generative models. Emphasis is placed on applying these methods to representative physics problems and on understanding the relationship between model behaviour and underlying physical assumptions.
Unlock more information by signing in with your HKUST account.