Seminari

Generative Adversarial Networks: a review of possible applications

Name: Generative Adversarial Networks: a review of possible applications
Start: 2022-12-15T14:30:00+01:00
End: 2022-12-15T15:30:00+01:00
Location: No location set

by Matteo Lai (Dipartimento di Ingegneria dell'Energia Elettrica e dell'Informazione - Università di Bologna)

Thursday 15 Dec 2022, 14:30 → 15:30 Europe/Rome

Sala Venturi

Description

Generative Adversarial Networks (GANs) are a class of machine learning frameworks made up of two neural networks which are trained in a zero-sum game, pitting one against the other, in order to generate synthetic instances of data. We will discuss the evolution of GANs since they were first proposed by Goodfellow et al. in 2014, describing their architecture and some alternatives of loss function to be used for training. Several extensions will be analysed, such as the conditional GAN, Auxiliary Classifier GAN (ACGAN), InfoGAN, Pix2Pix and StackGAN. To conclude, some advanced versions will be presented: CycleGAN, Progressive GAN and StyleGAN.