THE FUNDAMENTALS OF DEEP LEARNING

Content

Weeks	Topics
1. Weeks	The course begins with an introduction to the concepts of artificial intelligence, machine learning, and deep learning, explaining the differences between them, their historical development, and their place in the context of big data.
2. Weeks	The fundamentals of artificial neural networks, the artificial neuron model, activation functions, and simple feedforward networks are covered.
3. Weeks	Covers how networks are trained using forward propagation, error functions, and the backpropagation algorithm.
4. Weeks	Moves on to optimization methods; algorithms such as SGD, Momentum, RMSProp, and Adam are examined with examples.
5. Weeks	Addresses the issue of overfitting; regularization techniques, dropout, and data augmentation methods are covered.
6. Weeks	Deep learning libraries used in the Python environment (TensorFlow, PyTorch) are introduced, and a small neural network application is implemented.
7. Weeks	Covers multi-layer deep neural networks (DNN) in detail and applies them to simple classification problems.
8. Weeks	The logic of convolutional neural networks (CNN), layer structures, and their applications in image processing.
9. Weeks	Recurrent neural networks (RNN); their use in sequential data such as time series and text is explained alongside LSTM and GRU models.
10. Weeks	Natural language processing applications; embedding methods, Word2Vec, and basic text classification studies are conducted.
11. Weeks	Introduces encoder–decoder architectures and seq2seq models, and provides an introduction to the attention mechanism.
12. Weeks	Focuses on modern approaches; the Transformer architecture is examined in detail, and the fundamental principles of models such as BERT and GPT are discussed.
13. Weeks	Covers different application areas of deep learning (image, text, and audio processing) and its integration with big data ecosystems.
14. Weeks	Features presentations of term projects by students, a summary of topics covered in class, and an evaluation of future trends in deep learning.