Abstract
Inference algorithms for topic models are typ-
ically designed to be run over an entire col-
lection of documents after all of the docu-
ments have been observed. However, in many
applications of these models, the collection
of documents grows in size over time, mak-
ing it infeasible to run batch algorithms re-
peatedly. This problem can be addressed by
using online inference algorithms, which up-
date estimates of the topics as each document
is observed. We introduce two related Rao-
Blackwellized online inference algorithms for
the latent Dirichlet allocation (LDA) model
incremental Gibbs samplers and particle l-
ters and compare their runtime and perfor-
mance to that of existing algorithms.
Users
Please
log in to take part in the discussion (add own reviews or comments).