Generative Deep Learning

General Information
Teacher: Danilo Comminiello
Credits: 3 CFU
Scientific sector: ING-IND/31
Course language: English
Offered programs: PhD Course in Information and Communication Technology (ICT).
Calendar: July 4-5-11-14-17-19, 2023
Class timing: 10:00-13:00
Lecture modality: Hybrid:  DIET Reading Room (DIET Dept, II floor, Via Eudossiana 18) / Online via Zoom
Classroom page: Download slides, videolectures, notebooks, and additional material. Access code: 76btjj3. Participants are invited to register here.
GitHub repository: Here you can access a GitHub repository of the course with Python notebooks, additional material, project notebooks.
Office hours: by appointment

Course Description
Generative deep learning represents one of the most promising paradigms of modern artificial intelligence. Generative models aim at learning the true data distribution of the training set in an unsupervised fashion and at generating new data points with some variations by leveraging the capabilities of deep neural networks. The potential of these models is shown by some recent impressive tools, like Chat GPT, Midjourney, Stable Diffusion, or DALL-2.

In this course, we will study the foundations and the main models of generative deep learning, including variational autoencoders, generative adversarial networks, normalizing flows, energy-based models and diffusion models. The course will also discuss some applications related to information and communication technology (ICT) that benefit from deep generative learning. Lab sessions will be carried out using Python. 

Application examples can be agreed upon with the teacher based on their own PhD research topics. Please inform the teacher in advance.

TensorFlow Faculty Award
The course of Generative Deep Learning has been awarded the TensorFlow Faculty Award 2021 from Google in support of the development of new teaching courses on emerging machine learning topics that also promote diversity initiatives aimed at widening access to machine learning education.

Prerequisites
Knowledge of machine learning is warmly recommended.

Program

• INTRODUCTION TO GENERATIVE DEEP LEARNING. Generative modeling and probabilistic generative models. Deep neural networks. Building deep network models. Learning latent representations. 

• GENERATIVE AUTOREGRESSIVE AND FLOW-BASED MODELS. Autoregressive models. Parameterization by neural networks. Generation examples. Flows for continuous and discrete random variables. Invertible transformation of complex distributions. Normalizing flows. 

• VARIATIONAL AUTOENCODERS. Probabilistic principal component analysis. Variational inference. Variational autoencoders (VAEs) and modern deep VAE architectures.

• GENERATIVE ADVERSARIAL NETWORKS. Generator and critic. Training generative adversarial networks (GANs). Optimization and loss functions of GANs.

• DIFFUSION MODELS. Diffusion models. Score-based models. Advanced architectures and applications.

• APPLICATIONS. Python examples of generative deep learning models, including: music generation, image style transfer, text generation, video generation, anomaly detection, data augmentation, inverse problems solution, semantic communications, and medical imaging applications, among others.

Course flyer 

Class Schedule
The course will be held in July 2023 with the following schedule:

• Tuesday, July 4 10:00-13:00

• Wednesday, July 5 10:00-13:00

• Tuesday, July 11 10:00-13:00

• Friday, July 14  10:00-13:00

• Monday, July 17 10:00-13:00

• Wednesday, July 19 10:00-13:00

The course will be held in a hybrid modality:

• In-person: DIET Reading Room, DIET Dept., II floor, Via Eudossiana 18

• Online: via Zoom at the following link:
  https://uniroma1.zoom.us/j/86596351314?pwd=T0pja1RKLzFTd2ZabWN0TS9yN2tMQT09

Final Examination
Small project assignment on one of the course topics. The topic can also be agreed based on the students' PhD research program.

Research ideas can be developed by group of students.

Textbooks and Material

Course slides, lecture notes, and lab notebooks by the instructor.

All the material can be found on the Classroom page of the course. Participants are invited to register.

Textbooks:

• Jakub M. Tomczak, "Deep Generative Modeling", Springer, 2022.

• Kevin P. Murphy, "Probabilistic Machine Learning: Advanced Topics", The MIT Press, 2022.

• David Foster, "Generative Deep Learning: Teaching Machines to Paint, Write, Compose, and Play", O'Reilly Media, Inc., June 2019.