Modeling annotated data

DMCA

Modeling annotated data (2003)

Cached

Download Links

Other Repositories/Bibliography

DBLP

by David M. Blei , Michael I. Jordan

Venue:	IN PROC. OF THE 26TH INTL. ACM SIGIR CONFERENCE
Citations:	443 - 12 self

BibTeX

@INPROCEEDINGS{Blei03modelingannotated,
    author = {David M. Blei and Michael I. Jordan},
    title = {Modeling annotated data},
    booktitle = {IN PROC. OF THE 26TH INTL. ACM SIGIR CONFERENCE},
    year = {2003},
    publisher = {}
}

OpenURL

Abstract

We consider the problem of modeling annotated data—data with multiple types where the instance of one type (such as a caption) serves as a description of the other type (such as an image). We describe three hierarchical probabilistic mixture models that are aimed at such data, culminating in the Corr-LDA model, a latent variable model that is effective at modeling the joint distribution of both types and the conditional distribution of the annotation given the primary type. We take an empirical Bayes approach to finding parameter estimates and conduct experiments in held-out likelihood, automatic annotation, and text-based image retrieval using the Corel database of images and captions.

Keyphrases

latent variable model multiple type automatic annotation corr-lda model held-out likelihood primary type conditional distribution joint distribution text-based image retrieval hierarchical probabilistic mixture model data data parameter estimate empirical bayes approach corel database conduct experiment

DMCA