This course aims to cover advanced machine learning topics such as managing complex data, neural network architectures, and physics-informed neural networks. Students will develop expertise in applying these techniques to real-world problems, with critical evaluation of performance, and effective result interpretation. Furthermore, the course aims to provide students with the technical required to apply machine learning across various data science applications.
1. Advanced ML techniques (1.1 Scope and limitations of conventional ML models, 1.2 Ensemble learning Bagging, boosting and random forests, 1.3 Stacking and Blending)
2. Deep learning (2.1 Multilayer perceptrons layers, activation functions, loss, 2.2 Backpropagation and alternative methods Adam, RMSprop, AdaGrad, 2.3 Convolutional neural networks image classification, object detection)
3. Sequence-to-sequence neural networks (3.1 Recurrent architectures RNN, LSTM, 3.2 Transformers)
4. Representation Learning and generative models (4.1 Autoencoders denoising, clustering, 4.2 Variational autoencoders distribution learning, 4.3 Generative Adversarial networks image generation, inpainting)
5. Advanced topics in deep learning (5.1 Transfer learning and meta-learning, 5.2 Explainability and Ethics in AI, 5.3 State of the art models Graph neural networks, physics-informed models, equivariant neural networks)
By the end of the course, students should be able to do the following:
Curriculum explorer: Click here
SCQF Level: 10
Credits: 15