Online Activity Inference with GMTK
Jeffrey A. Bilmes, Richard Rogers

Citation
Jeffrey A. Bilmes, Richard Rogers. "Online Activity Inference with GMTK". Talk or presentation, 29, October, 2014.

Abstract
The Graphical Models Toolkit (GMTK) is an open source, publicly available toolkit for rapidly prototyping statistical models using dynamic graphical models (DGMs) and dynamic Bayesian networks (DBNs). GMTK can be used for any time-series application, such as real-time activity recognition. GMTK has many features, including exact and approximate inference; a large variety of built-in factors including dense, sparse, and deterministic conditional probability tables, native support for ARPA backoff-based factors and factored language models, parameter sharing, gamma and beta distributions, dense and sparse Gaussian factors, heterogeneous mixtures, deep neural network factors, and time-inhomogeneous trellis factors; arbitrary order embedded Markov chains; a GUI-based graph viewer; flexible feature-file support and processing tools (supporting pfiles, HTK files, ASCII/binary, and HDF5 files); and both offline and streaming online inference methods that can be used for both parameter learning and prediction. In this poster, we will outline how GMTK can be useful for designing and using models for real-time inference of activities based on multiple streams of sensor information as in a Terraswarm environment.

Electronic downloads

Internal. This publication has been marked by the author for TerraSwarm-only distribution, so electronic downloads are not available without logging in.

Citation formats

• HTML

  Jeffrey A. Bilmes, Richard Rogers. <a
  href="http://www.terraswarm.org/pubs/404.html"
  ><i>Online Activity Inference with
  GMTK</i></a>, Talk or presentation,  29,
  October, 2014.

• Plain text

  Jeffrey A. Bilmes, Richard Rogers. "Online Activity
  Inference with GMTK". Talk or presentation,  29,
  October, 2014.

• BibTeX

  @presentation{BilmesRogers14_OnlineActivityInferenceWithGMTK,
      author = {Jeffrey A. Bilmes and Richard Rogers},
      title = {Online Activity Inference with GMTK},
      day = {29},
      month = {October},
      year = {2014},
      abstract = {The Graphical Models Toolkit (GMTK) is an open
                source, publicly available toolkit for rapidly
                prototyping statistical models using dynamic
                graphical models (DGMs) and dynamic Bayesian
                networks (DBNs). GMTK can be used for any
                time-series application, such as real-time
                activity recognition. GMTK has many features,
                including exact and approximate inference; a large
                variety of built-in factors including dense,
                sparse, and deterministic conditional probability
                tables, native support for ARPA backoff-based
                factors and factored language models, parameter
                sharing, gamma and beta distributions, dense and
                sparse Gaussian factors, heterogeneous mixtures,
                deep neural network factors, and
                time-inhomogeneous trellis factors; arbitrary
                order embedded Markov chains; a GUI-based graph
                viewer; flexible feature-file support and
                processing tools (supporting pfiles, HTK files,
                ASCII/binary, and HDF5 files); and both offline
                and streaming online inference methods that can be
                used for both parameter learning and prediction.
                In this poster, we will outline how GMTK can be
                useful for designing and using models for
                real-time inference of activities based on
                multiple streams of sensor information as in a
                Terraswarm environment. },
      URL = {http://terraswarm.org/pubs/404.html}
  }
  

Posted by Jeffrey A. Bilmes on 28 Oct 2014.
Groups: tools