- Docente: Stefano Pagliarani
- Credits: 6
- SSD: MAT/06
- Language: English
- Moduli: Stefano Pagliarani (Modulo 1) Giovanni Paolini (Modulo 2)
- Teaching Mode: Traditional lectures (Modulo 1) Traditional lectures (Modulo 2)
- Campus: Bologna
-
Corso:
First cycle degree programme (L) in
Mathematics (cod. 8010)
Also valid for Second cycle degree programme (LM) in Mathematics (cod. 5827)
-
from Feb 19, 2024 to May 29, 2024
Learning outcomes
At the end of the course the students are familiar with the classical mathematical pillars of machine learning, including probabilistic aspects of advanced models such as neural networks and stochastic optimization algorithms for their training.
Course contents
- Introduction to machine learning. Regression and classification problems. Supervised, unsupervised and reinforcement learning. Overfitting and regularization.
- Foundations of information theory.
- Linear models for supervised learning. Support Vector Machine (SVM).
- Unsupervised learning: clustering, latent models, matrix factorization.
- Probabilistic models.
- Neural networks, backpropagation, stochastic gradient descent.
- Deep learning, convolutional networks, transformer.
- Selection of advanced topics: diffusion models, large language models, AlphaGo, AlphaTensor.
Readings/Bibliography
- Christopher Bishop, Pattern Recognition and Machine Learning
- Aston Zhang, Zack C. Lipton, Mu Li, Alex J. Smola, Dive into Deep Learning
Teaching methods
- Frontal lectures on the board and/or with slides.
- Coding and simulation activities in the laboratory.
Assessment methods
Oral examination and, possibly, presentation of one or more projects.
Teaching tools
- Office hours and tutoring.
- PDF lecture notes covering some parts of the program.
Office hours
See the website of Stefano Pagliarani
See the website of Giovanni Paolini