University of Cambridge > Talks.cam > Inference Group > Epitomes And Epitexts

Epitomes And Epitexts

Add to your list(s) Download to your calendar using vCal

If you have a question about this talk, please contact Phil Cowans.

The epitome model of images, originally proposed by Jojic and Frey divides the image into overlapping patches, each of which is generated probabilistically from a smaller image known as the epitome. The epitome therefore represents local texture in the original image. Applications of this model include reconstrustion of occluded pixels in video and image super-resolution.

In this talk I will describe our application of the same technique to text modelling. For this task, fragments of the text being modelled are drawn from distributions defined by the equivalent of the epitome, the epitext. This can be viewed as a generalisation of the ‘topics’ used in models such as latent Dirichlet allocation.

This is joint work with John Winn.

This talk is part of the Inference Group series.

Tell a friend about this talk:

This talk is included in these lists:

Note that ex-directory lists are not shown.

 

© 2006-2019 Talks.cam, University of Cambridge. Contact Us | Help and Documentation | Privacy and Publicity