Gaussian Process Topic Models

Amrudin Agovic, Arindam Banerjee
Proceedings of the 26th Conference on Uncertainty in Artificial Intelligence, PMLR R8:10-19, 2010.

Abstract

We introduce Gaussian Process Topic Mod- els (GPTMs), a new family of topic mod- els which can leverage a kernel among doc- uments while extracting correlated topics. GPTMs can be considered a systematic gen- eralization of the Correlated Topic Models (CTMs) using ideas from Gaussian Process (GP) based embedding. Since GPTMs work with both a topic covariance matrix and a document kernel matrix, learning GPTMs involves a novel component—solving a suit- able Sylvester equation capturing both topic and document dependencies. The efficacy of GPTMs is demonstrated with experiments evaluating the quality of both topic model- ing and embedding.

Cite this Paper

BibTeX
@InProceedings{pmlr-vR8-agovic10a, title = {{G}aussian Process Topic Models}, author = {Agovic, Amrudin and Banerjee, Arindam}, booktitle = {Proceedings of the 26th Conference on Uncertainty in Artificial Intelligence}, pages = {10--19}, year = {2010}, editor = {Grünwald, Peter and Spirtes, Peter}, volume = {R8}, series = {Proceedings of Machine Learning Research}, month = {08--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/r8/main/assets/agovic10a/agovic10a.pdf}, url = {https://proceedings.mlr.press/r8/agovic10a.html}, abstract = {We introduce Gaussian Process Topic Mod- els (GPTMs), a new family of topic mod- els which can leverage a kernel among doc- uments while extracting correlated topics. GPTMs can be considered a systematic gen- eralization of the Correlated Topic Models (CTMs) using ideas from Gaussian Process (GP) based embedding. Since GPTMs work with both a topic covariance matrix and a document kernel matrix, learning GPTMs involves a novel component—solving a suit- able Sylvester equation capturing both topic and document dependencies. The efficacy of GPTMs is demonstrated with experiments evaluating the quality of both topic model- ing and embedding.}, note = {Reissued by PMLR on 04 October 2026.} }
Endnote
%0 Conference Paper %T Gaussian Process Topic Models %A Amrudin Agovic %A Arindam Banerjee %B Proceedings of the 26th Conference on Uncertainty in Artificial Intelligence %C Proceedings of Machine Learning Research %D 2010 %E Peter Grünwald %E Peter Spirtes %F pmlr-vR8-agovic10a %I PMLR %P 10--19 %U https://proceedings.mlr.press/r8/agovic10a.html %V R8 %X We introduce Gaussian Process Topic Mod- els (GPTMs), a new family of topic mod- els which can leverage a kernel among doc- uments while extracting correlated topics. GPTMs can be considered a systematic gen- eralization of the Correlated Topic Models (CTMs) using ideas from Gaussian Process (GP) based embedding. Since GPTMs work with both a topic covariance matrix and a document kernel matrix, learning GPTMs involves a novel component—solving a suit- able Sylvester equation capturing both topic and document dependencies. The efficacy of GPTMs is demonstrated with experiments evaluating the quality of both topic model- ing and embedding. %Z Reissued by PMLR on 04 October 2026.
APA
Agovic, A. & Banerjee, A.. (2010). Gaussian Process Topic Models. Proceedings of the 26th Conference on Uncertainty in Artificial Intelligence, in Proceedings of Machine Learning Research R8:10-19 Available from https://proceedings.mlr.press/r8/agovic10a.html. Reissued by PMLR on 04 October 2026.

Related Material