Empirical Inference Conference Paper 2007

Bayesian Inference for Sparse Generalized Linear Models

We present a framework for efficient, accurate approximate Bayesian inference in generalized linear models (GLMs), based on the expectation propagation (EP) technique. The parameters can be endowed with a factorizing prior distribution, encoding properties such as sparsity or non-negativity. The central role of posterior log-concavity in Bayesian GLMs is emphasized and related to stability issues in EP. In particular, we use our technique to infer the parameters of a point process model for neuronal spiking data from multiple electrodes, demonstrating significantly superior predictive performance when a sparsity assumption is enforced via a Laplace prior distribution.

Author(s): Seeger, M. and Gerwinn, S. and Bethge, M.
Book Title: ECML 2007
Journal: Machine Learning: ECML 2007
Pages: 298-309
Year: 2007
Month: September
Day: 0
Series: Lecture Notes in Computer Science ; 4701
Editors: Kok, J. N., J. Koronacki, R. Lopez de Mantaras, S. Matwin, D. Mladenic, A. Skowron
Publisher: Springer
Bibtex Type: Conference Paper (inproceedings)
Address: Berlin, Germany
DOI: 10.1007/978-3-540-74958-5_29
Event Name: 18th European Conference on Machine Learning
Event Place: Warsaw, Poland
Digital: 0
Electronic Archiving: grant_archive
Language: en
Organization: Max-Planck-Gesellschaft
School: Biologische Kybernetik
Links:

BibTex

@inproceedings{4807,
  title = {Bayesian Inference for Sparse Generalized Linear Models},
  journal = {Machine Learning: ECML 2007},
  booktitle = {ECML 2007},
  abstract = {We present a framework for efficient, accurate approximate Bayesian inference in generalized linear models (GLMs), based on the expectation propagation (EP) technique. The parameters can be endowed with a factorizing prior distribution, encoding properties such as sparsity or non-negativity. The central role of posterior log-concavity in Bayesian GLMs is emphasized and related to stability issues in EP. In particular, we use our technique to infer the parameters of a point process model for neuronal spiking data from multiple electrodes, demonstrating significantly superior predictive performance when a sparsity assumption is enforced via a Laplace prior distribution.},
  pages = {298-309},
  series = {Lecture Notes in Computer Science ; 4701},
  editors = {Kok, J. N., J. Koronacki, R. Lopez de Mantaras, S. Matwin, D. Mladenic, A. Skowron},
  publisher = {Springer},
  organization = {Max-Planck-Gesellschaft},
  school = {Biologische Kybernetik},
  address = {Berlin, Germany},
  month = sep,
  year = {2007},
  slug = {4807},
  author = {Seeger, M. and Gerwinn, S. and Bethge, M.},
  month_numeric = {9}
}