Vai al contenuto principale

Seminario di Comunicazione Scientifica | Deep generative neural networks and an application in gene expression analysis

Pubblicato: Giovedì 10 ottobre 2024

Annuncio Seminario

Martedì 22 ottobre ore 17 Aula C
Speaker: Anders Krogh -Department of Computer Science and Center for Health Data Science, University of Copenhagen

Titolo:
Deep generative neural networks and an application in gene expression analysis

Abstract:
Deep generative neural networks are transforming society and  we already see a significant impact in biomedical research. I will give a brief introduction to the field and then focus on two simple types of models in which a neural network maps from a low dimensional
representation in a latent space to the samples in feature space.
One is the Variational Autoencoder (VAE), which uses variational inference, and
the other is the encoder-less deep generative decoder (DGD), wich uses maximum likelihood to estimate representations. In the second half of the talk, I will present a couple of applications of a DGD for modeling of gene expression data.

Link Webex: https://unito.webex.com/unito/j.php?MTID=m46a93056764c6dd53eb268b4195e4ae2

L'incontro rientra nell'ambito del credito per il Seminario di Comunicazione Scientifica 

Relativo ai corsi di:
Ultimo aggiornamento: 22/10/2024 10:24
Location: https://fisica-sc.campusnet.unito.it/robots.html
Non cliccare qui!