The Infinite Latent Events Model

• Main Authors: Wingate, David (Contributor), Goodman, Noah D. (Contributor), Roy, Daniel (Contributor), Tenenbaum, Joshua B. (Contributor)
• Other Authors: Massachusetts Institute of Technology. Department of Brain and Cognitive Sciences (Contributor), Massachusetts Institute of Technology. Laboratory for Information and Decision Systems (Contributor)
• Format: Article
• Language: English
• Published: Association for Uncertainty in Artificial Intelligence Press, 2012-06-28T15:56:37Z.
• Subjects: Article
• Online Access: Get fulltext

We present the Infinite Latent Events Model, a nonparametric hierarchical Bayesian distribution over infinite dimensional Dynamic Bayesian Networks with binary state representations and noisy-OR-like transitions. The distribution can be used to learn structure in discrete timeseries data by simultaneously inferring a set of latent events, which events fired at each timestep, and how those events are causally linked. We illustrate the model on a sound factorization task, a network topology identification task, and a video game task.
NTT Communication Science Laboratories
United States. Air Force Office of Scientific Research (AFOSR FA9550-07-1-0075)
United States. Office of Naval Research (ONR N00014-07-1-0937)
National Science Foundation (U.S.) (Graduate Research Fellowship)
United States. Army Research Office (ARO W911NF-08-1-0242)
James S. McDonnell Foundation (Causal Learning Collaborative Initiative)