Bims


Bayesian Inference of Model Structure.

Bims constructs Markov chains over model structures defined by Dlps (Distributional Logic Programming programs).

The main intuition is that a DLP program can define the space of all possible models and also assign a probability which reflects our prior belief that a particular model explains the data. In addition to composing such a program, the user must provide a function for calculating the likelihood of the data given a model.

install

SWI

Bims can be installed from within SWI using its package manager.

?- pack_install(bims).

And then load by :

?- [library(bims)].

To install a source distribution:

?- pack_install( 'the_source.tgz' ).

Yap

Copy the sources from bims
and install locally. To use load the file

?- ['prolog/bims.pl']

examples

There are 2 main prepared examples.

?- bims([]).

Runs 3 short example chains that learn classifications trees for the BCW dataset.


?- bims([models(bns)]).

Runs chains that learn Bayesian networks.

docs

Also available in distribution directory doc/

sources

packed sources: bims
sources on github bims

authors

Nicos Angelopoulos
James Cussens

---

Invited talk

Invited talk at PLP 2015 (September 2015, Cork, Ireland)

Bayesian inference of model structure

Publications

When citing Bims use the following:
Nicos Angelopoulos and James Cussens. Distributional Logic Programming for Bayesian Knowledge Representation.
International Journal of Approximate Reasoning (IJAR). To appear in Volume 80, January 2017, pages 52-66.
Available on-line August 26th, 2016.

Papers using Bims
Nicos Angelopoulos and Lodewyk Wessels. Effective priors over model structures applied to DNA binding assay data .
AIME'11 Workshop on Probabilistic Problem Solving in Biomedicine Bled, Slovenia, July 2011.

Nicos Angelopoulos, Andreas Hadjiprocopis and Malcolm D. Walkinshaw. Learning binding affinity from augmented high throuput screening data. Chapter in "Chemoinformatics and Advanced Machine Learning Perspectives: Complex Computational Methods and Collaborative Techniques". IGI Book Series on Chemoinformatics. Huma Lodhi and Yoshihiro Yamanishi (Eds). 212-234, 2011.

Nicos Angelopoulos, Andreas Hadjiprocopis and Malcolm D. Walkinshaw. Bayesian ligand discovery from high dimensional descriptor data. ACS Journal of Chemical Information and Modeling , 49(6), 1547-1557, 2009.
Earlier papers explaining technical aspects of Bims

Nicos Angelopoulos and James Cussens Bayesian learning of Bayesian Networks with informative priors. Special issue on BN learning. Journal of Annals of Mathematics and Artificial Intelligence , 54(1-3), 53-98, 2008.

Nicos Angelopoulos and James Cussens. Tempering for Bayesian C&RTs. Proceedings of the 2nd International Conference on Machine Learning (ICML05), Bonn, 2005.

Nicos Angelopoulos and James Cussens. Exploiting Informative Priors for Bayesian Classification and Regression Trees. Proceedings of the Nineteenth International Joint Conference on Artificial Intelligence , Edinburgh, Scotland, UK, July - August 2005.

Nicos Angelopoulos and James Cussens. Extended stochastic logic programs for informative priors over C&RTs. In Rui Camacho, Ross King, and Ashwin Srinivasan, editors, Proceedings of the work-in-progress track of the Fourteenth International Conference on Inductive Logic Programming (ILP04), pages 7-11, Porto, September 2004.

Nicos Angelopoulos and James Cussens. On the implementation of MCMC proposals over stochastic logic programs. In Colloquium on Implementation of Constraint and LOgic Programming Systems. Satellite workshop to ICLP'04, Saint-Malo, France, 2004.

Nicos Angelopoulos and James Cussens. Markov chain Monte Carlo using tree-based priors on model structure. In Jack Breese and Daphne Koller, editors, Proceedings of the Seventeenth Annual Conference on Uncertainty in Artificial Intelligence (UAI-2001), Seattle, August 2001. Morgan Kaufmann.


Nicos Angelopoulos,
London, April, 2017