Accelerated sampling for the Indian Buffet Process Export

@inproceedings{citeulike:5747702, abstract = {We often seek to identify co-occurring hidden features in a set of observations. The Indian Buffet Process (IBP) provides a non-parametric prior on the features present in each observation, but current inference techniques for the IBP often scale poorly. The collapsed Gibbs sampler for the IBP has a running time cubic in the number of observations, and the uncollapsed Gibbs sampler, while linear, is often slow to mix. We present a new linear-time collapsed Gibbs sampler for conjugate likelihood models and demonstrate its efficacy on large real-world datasets.}, address = {New York, NY, USA}, author = {Velez, Finale D. and Ghahramani, Zoubin}, booktitle = {ICML '09: Proceedings of the 26th Annual International Conference on Machine Learning}, citeulike-article-id = {5747702}, citeulike-linkout-0 = {http://portal.acm.org/citation.cfm?id=1553374.1553409}, citeulike-linkout-1 = {http://dx.doi.org/10.1145/1553374.1553409}, doi = {10.1145/1553374.1553409}, isbn = {978-1-60558-516-1}, keywords = {bayesian, gibbs, indian-buffet-process, inference, nonparametric, prior, process, sampling}, location = {Montreal, Quebec, Canada}, pages = {273--280}, posted-at = {2009-09-08 20:35:56}, priority = {2}, publisher = {ACM}, title = {Accelerated sampling for the Indian Buffet Process}, url = {http://dx.doi.org/10.1145/1553374.1553409}, year = {2009} }

TY - CONF ID - citeulike:5747702 L3 - citeulike-article-id:5747702 N2 - We often seek to identify co-occurring hidden features in a set of observations. The Indian Buffet Process (IBP) provides a non-parametric prior on the features present in each observation, but current inference techniques for the IBP often scale poorly. The collapsed Gibbs sampler for the IBP has a running time cubic in the number of observations, and the uncollapsed Gibbs sampler, while linear, is often slow to mix. We present a new linear-time collapsed Gibbs sampler for conjugate likelihood models and demonstrate its efficacy on large real-world datasets. CY - Montreal, Quebec, Canada SN - 978-1-60558-516-1 AD - New York, NY, USA EP - 280 TI - Accelerated sampling for the Indian Buffet Process PB - ACM T2 - ICML '09: Proceedings of the 26th Annual International Conference on Machine Learning SP - 273 KW - bayesian KW - gibbs KW - indian-buffet-process KW - inference KW - nonparametric KW - prior KW - process KW - sampling AU - Velez, Finale AU - Ghahramani, Zoubin PY - 2009/// UR - http://dx.doi.org/10.1145/1553374.1553409 ER -