Guest editorial
Probabilistic problem solving in biomedicine

https://doi.org/10.1016/j.artmed.2013.01.006Get rights and content

Introduction

With the current trend towards pervasive health care (see e.g. [1]), personalised health care (see e.g. [2]), and the ever growing amount of evidence coming from biomedical research, methods that can handle reasoning and learning under uncertainty are becoming more and more important. Probabilistic methods, and in particular Bayesian networks (BNs), have been introduced in the 1980s as a formalism for representing and reasoning with models of problems involving uncertainty, adopting probability theory as a basic framework. Since the beginning of the 1990s, researchers are exploring its possibilities for developing medical applications. An early example is the Pathfinder project, where a Bayesian network was developed for providing assistance with the identification of disorders from lymph-node tissue sections [3], which was later commercialised.

The ongoing developments of the past two decades in the field of artificial intelligence have made it now possible to apply probabilistic methods to – in principle – solve problems in real-world biomedical domains. However, the practice of applying new techniques to real-world medical problem leads to new interesting problems that are often not considered in computer science research. The five papers in this special are typical examples of papers which address issues of applying probabilistic methods to medicine in different ways. The papers in this special issues are based on a selection of papers originally submitted to the workshop Probabilistic Problem Solving in Biomedicine (ProBioMed), held in Bled, Slovania, in conjunction with the 13th Conference on Artificial Intelligence in Medicine (AIME’11).

The authors of the papers which had the best review scores and were considered to be most suitable for this journal, the international journal Artificial Intelligence in Medicine (AIIM), were invited to submit a revised and extended version for review and inclusion for a special AIIM issue. After the usual revision process, the resulting five papers are now presented in this current special issue of AIIM.

In this editorial we sketch the scientific context of probabilistic problem solving and we briefly introduce the selected papers.

Section snippets

Dealing with complexity

Biomedical data and knowledge are usually complex and part of the work in data manipulation, knowledge elicitation and modelling concerns finding solutions that allow dealing with, and sometimes reducing, this complexity. This is also visible in the contributions to this special issue.

Conclusions

This special issue of Artificial Intelligence in Medicine contains five papers which deal with the complexities of problem solving based on probabilistic models, in particular graphical models. The papers in this special issue each propose different approaches to do so. We think that this shows that the important research issues shifted somewhat from the question whether we can actually learn and build graphical models for biomedical problems; the real challenges are related to the question how

Acknowledgements

We would like to thank the program committee members of the ProBioMed workshop and in particular the people who were involved in reviewing the papers of this special issue: Hendrik Blockeel, Carlo Combi, Luis M. de Campos, Javier Díez, Milos Hauskrecht, Jesse Hoey, Pedro Larrañaga, William Marsh, Stijn Meganck, Agnieszka Oniśko, Niels Peek, Alberto Riva, Enrique Sucar, and Allan Tucker. We are thankful to all of them for their devotion to aiming for a high quality special issue.

References (9)

There are more references available in the full text version of this article.

Cited by (0)

View full text
Copyright © 2013 Elsevier B.V. All rights reserved.