Reliable model evaluation is essential for developing trustworthy machine learning systems. This module introduces evaluation methodologies for different machine learning scenarios, including traditional model evaluation, sequential data evaluation, unstable model evaluation, imbalanced learning, and statistical comparison of machine learning models. Python implementation, assisted by an LLM, and practical case studies are incorporated throughout the module. After learning this module, students will be able to:

This module introduces the design principles and optimization techniques of neural networks, including weight initialization, optimization algorithms, normalization methods, regularization strategies, and data augmentation techniques. Python implementation using modern deep learning frameworks and LLM-assisted programming are integrated with practical engineering examples. After learning this module, students will be able to:

This module introduces modern deep learning architectures for processing images and sequential data, including convolutional neural networks (CNNs), recurrent neural networks (RNNs), long short-term memory (LSTM), gated recurrent units (GRU), attention mechanisms, and Transformers. Python implementation assisted by LLM and representative AI applications are covered throughout the module. After learning this module, students will be able to:

This module introduces the principles and applications of modern generative artificial intelligence, including traditional generative models, variational autoencoders (VAEs), normalizing flows, generative adversarial networks (GANs), and diffusion models. Python implementation assisted by LLM is combined with representative real-world case studies. After learning this module, students will be able to: