Continuous Time Dynamic Topic Models
This work addresses the need for more scalable and fine-grained topic modeling in text analysis, though it is incremental as it builds on existing dynamic topic models.
The paper tackles the problem of modeling evolving topics in document collections over time by developing a continuous time dynamic topic model (cDTM) that uses Brownian motion, resulting in efficient variational inference that handles many time points and outperforms discrete-time models in predictive perplexity and time stamp prediction tasks.
In this paper, we develop the continuous time dynamic topic model (cDTM). The cDTM is a dynamic topic model that uses Brownian motion to model the latent topics through a sequential collection of documents, where a "topic" is a pattern of word use that we expect to evolve over the course of the collection. We derive an efficient variational approximate inference algorithm that takes advantage of the sparsity of observations in text, a property that lets us easily handle many time points. In contrast to the cDTM, the original discrete-time dynamic topic model (dDTM) requires that time be discretized. Moreover, the complexity of variational inference for the dDTM grows quickly as time granularity increases, a drawback which limits fine-grained discretization. We demonstrate the cDTM on two news corpora, reporting both predictive perplexity and the novel task of time stamp prediction.