Stochastic Collapsed Variational Inference for Sequential Data
This work addresses the challenge of scalable inference for sequential data in machine learning, representing an incremental advancement by extending existing stochastic collapsed variational inference methods to sequential settings.
The paper tackles the problem of performing efficient and accurate inference for sequential data models, proposing a stochastic collapsed variational inference algorithm that is more efficient and accurate than its uncollapsed version, with experimental results on discrete datasets showing improvements.
Stochastic variational inference for collapsed models has recently been successfully applied to large scale topic modelling. In this paper, we propose a stochastic collapsed variational inference algorithm in the sequential data setting. Our algorithm is applicable to both finite hidden Markov models and hierarchical Dirichlet process hidden Markov models, and to any datasets generated by emission distributions in the exponential family. Our experiment results on two discrete datasets show that our inference is both more efficient and more accurate than its uncollapsed version, stochastic variational inference.