The paper introduces stochastic variational inference, a scalable algorithm for approximating posterior distributions in probabilistic models. The authors develop this technique for a wide range of probabilistic models and demonstrate its effectiveness on two topic models: latent Dirichlet allocation and the hierarchical Dirichlet process topic model. Using stochastic variational inference, they analyze large collections of documents, including 300K articles from *Nature*, 1.8M articles from *The New York Times*, and 3.8M articles from *Wikipedia*. The method outperforms traditional variational inference, which can only handle smaller subsets of such data sets. Stochastic variational inference allows complex Bayesian models to be applied to massive data sets, making it particularly useful for modern data analysis tasks that involve large, high-dimensional data. The paper also discusses the technical details of the algorithm, including the use of stochastic optimization and natural gradients, and provides a review of related work in variational inference and Markov chain Monte Carlo methods.The paper introduces stochastic variational inference, a scalable algorithm for approximating posterior distributions in probabilistic models. The authors develop this technique for a wide range of probabilistic models and demonstrate its effectiveness on two topic models: latent Dirichlet allocation and the hierarchical Dirichlet process topic model. Using stochastic variational inference, they analyze large collections of documents, including 300K articles from *Nature*, 1.8M articles from *The New York Times*, and 3.8M articles from *Wikipedia*. The method outperforms traditional variational inference, which can only handle smaller subsets of such data sets. Stochastic variational inference allows complex Bayesian models to be applied to massive data sets, making it particularly useful for modern data analysis tasks that involve large, high-dimensional data. The paper also discusses the technical details of the algorithm, including the use of stochastic optimization and natural gradients, and provides a review of related work in variational inference and Markov chain Monte Carlo methods.