PROGRAMMA DEL CORSO

Giorno 1

• Introduction to Artificial Intelligence
• AI, Machine learning and Deep Learning
• Examples of common use cases
• CRISP-DM methodology (Business understanding, Data understanding, Data preparation, Modeling, Evaluation, Deployment)
• Methodologies for managing the entire model lifecycle: ASUM-DM
• Preliminary exploratory data analisys techniques
• Most commonly used platform, framework and tools landscape
• Using Jypter notebooks
• Supervised learning algorithms (Linear and logistics regression, Decision trees, Scikit-learn library usage examples)

Giorno 2

• Ensable models: random forests
• K-nearest neighbors
• Suport vector machines
• Naives Bayes
• Using scikit-learn pipelines
• Unsupervised algorithms (K-means clustering, Density-based spatial clustering of applications with noise (DBSCAN)
• Model evalutaion (Regression metrics: MAE, MSE and R2 score, Metrics for classification: confusion matrix and ROC curve, Clustering evaluation strategies)
• Train/test split and k-fold cross validation
• Understanding overfitting and underfitting
• Dimensional reduction: PCS and LDA
• Hyperparameter optimization
• Tracking experiments: introducing MLFlow

Giorno 3

• Introduction to Artificial Neural Networks (Model of biological neuron, Activation function, SGD, Learning rate and momentum, Backpropagation)
• Most commonly used architectures and their uses (DFF, RNN, AE, MC, HN, DCN, GAN)
• Examples using Keras library
• Time series analysis (Terminology, Autocovariance and autocorrelations, Stationarity, ARIMA models)
• Neural networks for time serires analysis (Gated functions)
• Serving models (PMML and ONNX interchange formats)

Giorno 4

• Introduction to Big Data
• Understanding the 3Vs
• Horizontal vs. vertical scaling
• ACID, BASE, SEC and CAP theorem
• RDBMS vs. NoSQL vs. Big Data
• The Hadoop ecosystem
• From Data Warehouse to Data Lake
• Introducing Apache Spark
• Big Data in the Cloud era
• The market and the state-of-the-art technologies