Servizi per la didattica
PORTALE DELLA DIDATTICA

Adversarial training of neural networks

01UJBRV

A.A. 2019/20

Lingua dell'insegnamento

Inglese

Corsi di studio

Dottorato di ricerca in Ingegneria Elettrica, Elettronica E Delle Comunicazioni - Torino

Organizzazione dell'insegnamento
Didattica Ore
Lezioni 15
Docenti
Docente Qualifica Settore h.Lez h.Es h.Lab h.Tut Anni incarico
Valsesia Diego   Ricercatore L240/10 ING-INF/03 15 0 0 0 2
Collaboratori
Espandi

Didattica
SSD CFU Attivita' formative Ambiti disciplinari
*** N/A ***    
2019/20
PERIOD: MAY - JUNE Generative Adversarial Networks (GANs) have achieved impressive results in generating fake images that look realistic. GANs introduced the concept of adversarial training where two neural networks are trained by playing a game against each other. Adversarial training is a powerful concept that has been applied to wide variety of problems across different fields including generative models for images, time series, or DNA/protein sequences, unsupervised image-to-image translation, domain adaptation, regularization of inverse problems and many more. This course will provide students with the theoretical foundations of adversarial training as well as the most recent practical examples of its use. The target audience is cross-disciplinary as adversarial training has proved itself to be a staple of modern deep learning across many fields.
PERIOD: MAY - JUNE Generative Adversarial Networks (GANs) have achieved impressive results in generating fake images that look realistic. GANs introduced the concept of adversarial training where two neural networks are trained by playing a game against each other. Adversarial training is a powerful concept that has been applied to wide variety of problems across different fields including generative models for images, time series, or DNA/protein sequences, unsupervised image-to-image translation, domain adaptation, regularization of inverse problems and many more. This course will provide students with the theoretical foundations of adversarial training as well as the most recent practical examples of its use. The target audience is cross-disciplinary as adversarial training has proved itself to be a staple of modern deep learning across many fields.
• Brief review of neural network architectures and training via the backpropagation algorithm • Generative Adversarial Networks (GANs): introduction and game-theoretical formulation • Wasserstein GANs: stability improvements via optimal transport formulation • Generative models: recent advances in image generation (Progressive Growing, BigGAN), point cloud generation (GraphCNN), optimized DNA sequences • Image-to-image and video-to-video translation: style transfer and Cycle GANs • Inverse problems: image inpainting, superresolution, compressed sensing
• Brief review of neural network architectures and training via the backpropagation algorithm • Generative Adversarial Networks (GANs): introduction and game-theoretical formulation • Wasserstein GANs: stability improvements via optimal transport formulation • Generative models: recent advances in image generation (Progressive Growing, BigGAN), point cloud generation (GraphCNN), optimized DNA sequences • Image-to-image and video-to-video translation: style transfer and Cycle GANs • Inverse problems: image inpainting, superresolution, compressed sensing
Wednesday 11/3 9:00-12:00 Aula C Wednesday 18/3 14:00-17:00 Aula C Wednesday 25/3 14:00-17:00 Aula C Wednesday 1/4 14:00-17:00 Aula C Wednesday 8/4 14:00-17:00 Aula C ATTENZIONE: LE LEZIONI VERRANNO RIPROGRAMMATE PER IL PERIODO MAGGIO - GIUGNO
Wednesday 11/3 9:00-12:00 Aula C Wednesday 18/3 14:00-17:00 Aula C Wednesday 25/3 14:00-17:00 Aula C Wednesday 1/4 14:00-17:00 Aula C Wednesday 8/4 14:00-17:00 Aula C ATTENZIONE: LE LEZIONI VERRANNO RIPROGRAMMATE PER IL PERIODO MAGGIO - GIUGNO
Modalità di esame:
Exam:
Esporta Word


© Politecnico di Torino
Corso Duca degli Abruzzi, 24 - 10129 Torino, ITALY
Contatti