Machine Learning

, Volume 95, Issue 3, pp 423–469

Interactive topic modeling

  • Yuening Hu
  • Jordan Boyd-Graber
  • Brianna Satinoff
  • Alison Smith
Article

DOI: 10.1007/s10994-013-5413-0

Cite this article as:
Hu, Y., Boyd-Graber, J., Satinoff, B. et al. Mach Learn (2014) 95: 423. doi:10.1007/s10994-013-5413-0

Abstract

Topic models are a useful and ubiquitous tool for understanding large corpora. However, topic models are not perfect, and for many users in computational social science, digital humanities, and information studies—who are not machine learning experts—existing models and frameworks are often a “take it or leave it” proposition. This paper presents a mechanism for giving users a voice by encoding users’ feedback to topic models as correlations between words into a topic model. This framework, interactive topic modeling (itm), allows untrained users to encode their feedback easily and iteratively into the topic models. Because latency in interactive systems is crucial, we develop more efficient inference algorithms for tree-based topic models. We validate the framework both with simulated and real users.

Keywords

Topic models Latent Dirichlet Allocation Feedback Interactive topic modeling Online learning Gibbs sampling 

Copyright information

© The Author(s) 2013

Authors and Affiliations

  • Yuening Hu
    • 1
  • Jordan Boyd-Graber
    • 2
  • Brianna Satinoff
    • 1
  • Alison Smith
    • 1
  1. 1.Computer ScienceUniversity of MarylandCollege ParkUSA
  2. 2.iSchool and UMIACSUniversity of MarylandCollege ParkUSA

Personalised recommendations