The introductory course presents the topics underlying modern neural network architectures, paying particular attention to the mathematical foundations and their implications for the neural network landscape, to allow students to more fully understand what they are and how to catalog them. Through applied use of the technologies with which artificial intelligence tools are built, and a comparison with the design model with which these solutions are built, the goal is to enable students to interpret current and future scenarios.

Program

• Introduction to Google Colab
• ANN Framework
• Linear algebra calls
• Introduction to multidimensional arrays
• Elementary elements of an ANN
• Types of elementary operations on tensors
• Probability calculation calls
• Designing a neural network
• Types of training
• Most popular neural network models
• Attention mechanism
• Transformers
• GNN (Graph Neural Network), LLM and Retrieval-Augmented Generation RAG