Modeling annotated data
David M. Blei, Michael I. Jordan · 2003
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 which aim to describe such data, culminating in correspondence latent Dirichlet allocation, 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 conduct experiments on the Corel database of images and captions, assessing performance in terms of held-out likelihood, automatic annotation, and text-based image retrieval.
Cited by 1,054 works. doi:10.1145/860435.860460
Authors
References
- Latent dirichlet allocation
- Handbook of Mathematical Functions
- Normalized cuts and image segmentation
- An Introduction to Variational Methods for Graphical Models
- Parametric Empirical Bayes Inference: Theory and Applications
- A Language Modeling Approach to Information Retrieval
- A language modeling approach to information retrieval
- A Variational Baysian Framework for Graphical Models
- SIMPLIcity: semantics-sensitive integrated matching for picture libraries
- The Missing Link - A Probabilistic Model of Document Content and Hypertext Connectivity