Foundations of Deep Learning- Start of the Micro Degree

This is the micro degree program of the course series "Foundations of Deep Learning", comprising three courses, providing a comprehensive introduction to modern deep learning. In the part "Basics", learners begin by exploring core principles such as neural network architecture, backpropagation, optimisation, and regularisation. The part "Architectures & Methodology" delves into specialised models like CNNs, RNNs, and Transformers, while covering practical skills such as transfer learning and debugging. In the last part "Advanced Topics", topics such as generative models, uncertainty estimation, and hyperparameter optimisation are introduced. A micro degree certificate is issued upon completing all three courses. After clicking on the "Enrol/to course" button, the certificate can be found in the course section "Micro Degree Certificate." 

The three online courses from this micro degree program will cover:

• Core Concepts of Neural Networks: Learn the fundamentals of Multi-Layer Perceptrons (MLPs), the backpropagation algorithm, and key optimisation techniques for training deep networks.
• Regularisation and Model Robustness: Explore methods like dropout, data augmentation, and ensembling to prevent overfitting and improve generalisation.
• Specialised Architectures: Understand how Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), and Transformers are used for different types of data and tasks.
• Practical Deep Learning Skills: Gain hands-on knowledge in debugging, transfer learning, self-supervised learning, and real-world model development.
• Advanced Topics and Model Tuning: Delve into generative models (VAEs, GANs, diffusion models), uncertainty estimation, and hyperparameter optimisation (HPO) for high-performing deep learning systems.

Upon conclusion of this micro degree program, you will

• Understand the Fundamentals of Neural Networks, explain how core components such as MLPs, activation functions, loss functions, and backpropagation work, and how learning differs from optimisation.
• Apply and Compare Deep Learning Techniques, describe key optimisation algorithms (e.g., SGD, Adam), regularisation methods (e.g., dropout, early stopping), and normalisation techniques to improve training and generalization.
• Analyse and Design Advanced Neural Architectures, understand the structure and function of CNNs, RNNs (including LSTMs and GRUs), attention mechanisms, and the Transformer architecture.
• Implement Practical Deep Learning Strategies, apply concepts such as transfer learning, self-supervised learning, and debugging techniques to design robust and efficient deep learning systems.
• Explore Generative Models and Model Uncertainty, explain the mechanisms of autoencoders, GANs, diffusion models, and describe uncertainty estimation methods (e.g., Variational Inference, MCMC), as well as hyperparameter optimisation techniques.

To receive the micro degree, learners must have completed the three online courses of "Foundations of Deep Learning", i.e., Basics, Architectures & Methodology, Advanced Topics. The Transcript of Records is available to those who have achieved at least 60% of the total grade in all courses by completing the graded activities.