Scientific Programme

[Tutorial Programme] [Invited Speakers] [Accepted Papers] [Accepted Abstracts] [Accepted Posters]

Conference Schedule

For further information regarding the Conference Schedule, please follow this link: http://www.ai.univie.ac.at/aimdm99/schedule.html

Tutorial Programme

Half day tutorials will be organized on Sunday 20th June in parallel sessions. Morning tutorials will take place from 9 am to 12:30 pm and afternoon tutorials from 1:30 pm to 5 pm. Prices will be 400 DKK for one tutorial and 800 DKK for two. For registration, please follow this link. Proposals for tutorials are invited. Send proposals to the tutorials chair, Jeremy Wyatt (see Conference addresses) no later than December 15th 1998.

Please notice following: Tutorials 1, 4 and 5 runs in the morning and tutorials 2, 3 and 6 in the afternoon.


1) NATURAL LANGUAGE GENERATION

Audience: Researchers and software developers who are interested in improving the quality of documents and other written materials produced by software systems.
Description: Many medical IT systems need to produce documents or other types of written texts, such as discharge reports, letters to patients, and explanations of expert-system reasoning. The quality and readability of such texts is not always as high as it could be, unfortunately. This tutorial will discuss some of the linguistic problems that computer-generated texts can suffer from, such as poor rhetorical structure, inappropriate anaphors, false implicatures, and grammatical mistakes. Natural-language generation (NLG) technology is introduced, which can be used to automatically produce texts which satisfy linguistic constraints and hence do not suffer from these problems. The tutorial will be illustrated with examples from NLG systems developed at Aberdeen and elsewhere. Attendees do not need any background in linguistics or natural-language processing, but they should be familiar with basic AI concepts. Even people who do not intend to use NLG technology will still benefit from the tutorial, by becoming more aware of potential linguistic problems in computer-generated texts and how they can be resolved.

Tutorial Presenter:
Ehud Reiter
Dept of Computing Science, University of Aberdeen, Aberdeen AB24 3UE, UK.
Phone +44-1224-273443, Fax +44-1224-273422,
email: ereiter@csd.abdn.ac.uk


2) THE PSYCHOLOGY OF MEDICAL DECISION MAKING

This course is intended for clinicians, and others who wish to gain insight into the psychological factors influencing their decision making under uncertainty. The aim of this tutorial is to increase understanding of the psychological processes involved in medical decision making. This knowledge is useful in trying to improve decision making. Additionally it is a fruitful area for research. The course assumes the attendee has only a basic knowledge of the subject matter.
The clinician's reasoning is a partial cause of non-optimal medical decisions. The cognitive psychology of judgment and decision making offers explanations of how some of these reasoning errors are made. The course will review the basic nature of expert medical reasoning, to discover possibilities for capitalizing on its strengths and supporting its weaknesses. The participant will learn why the human cognitive system, with its large memory, limited attention span, and powerful pattern recognition ability, seems destined to operate by automatic "scripted" response rather than thoughtful deliberation.
We will demonstrate the implications of clinicians' cognitive processes for two basic activities of rational decision making: diagnosis and choosing a course of action. Clinicians' reasoning strategies, motivations, habits, and cognitive limitations can lead them to make errors of diagnosis, and define the methods they can use to seek and use information more rationally. Clinicians' strategies for predicting what will happen can lead to misjudgments of probability, and their methods of evaluating things can lead to misjudgments of treatment consequences. Understanding the psychological processes involved will suggest methods for helping clinicians reason better about the probabilities of outcomes and about their own or their patients' preferences.

Tutorial presenters:
Robert M. Hamm
Clinical Decision Making Program, Department of Family and Preventive Medicine,
University of Oklahoma Health Sciences Center,
900 NE 10th St, Oklahoma City, OK 73104, U.S.A.
Telephone: 405/271-8000 ext3-2302, Fax: 405/271-2784,
email: robert-hamm@ouhsc.edu

Clare Harries
Department of Psychology, University College London,
Gower Street,London, WC1E 6BT, U.K.
Telephone: +44 171 504 5389, Fax: +44 171 436 4276,
email: clare.harries@ucl.ac.uk

Jack Dowie, PhD
Department of Applied Social Sciences, The Open University,
Milton Keynes, MK7 6AA, U.K.
Telephone: +44 171 254 7576, Fax: +44 171 254 7576,
email: j.a.dowie@open.ac.uk


3) DATA MINING TECHNIQUES AND APPLICATIONS IN MEDICINE

With the widespread use of medical information systems that include databases which have recently featured explosive growth in their sizes, physicians and medical researchers are faced with a problem of making use of the stored data. The traditional manual data analysis has become insufficient, and methods for efficient computer-assisted analysis indispensable, in particular those of data mining and other related techniques of knowledge discovery in databases and intelligent data analysis.
This tutorial will address current techniques and applications of data mining in medicine. We will provide an overview of data mining methods, including symbolic data mining (mining of decision rules, association rules, decision trees, inductive logic programming, hierarchical concept discovery, etc.) and subsymbolic data mining (instance based learning, neural nets, Naive Bayesian classifier, etc). Specific evaluation techniques and statistical criteria suited for medical applications will be discussed. Selected data preprocessing and data visualization methods will also be presented.

The participants of tutorial will get familiar with:

Intended audience:
This tutorial will be of interest to clinicians, medical researchers, information technology professionals, information systems developers and managers, data analysts and institutional decision makers, and anyone else interested in applying modern data analysis methods to extract useful knowledge from medical data bases.

Tutorial presenters:
Blaz Zupan (1,2) and Nada Lavrac (2)
(1) University of Ljubljana,
Faculty of Computer and Information Sciences Trzaska 25,
SI-1000 Ljubljana, Slovenia.
Telephone: +386 61 177 3380, fax: +386 61 125 1038
email: blaz.zupan@fri.uni-lj.si

(2) J.Stefan Institute, Department of Intelligent Systems Jamova 39,
SI-1000 Ljubljana, Slovenia.
Telephone: +386 61 177 3272, fax: +386 61 125 1038.
email: nada.lavrac@ijs.si


4) HOW TO BUILD A CAUSAL PROBABILISTIC NETWORK

A Causal Probabilistic Network, also called Bayesian network is a flexible and efficient framework for reasoning under uncertainty, and it has established itself as a practical method for knowledge representation and inference in a number of medical areas. The framework consists of a structural part, where the domain in question is modelled through a directed acyclic graph, and a quantitative part, where the impact between nodes in the graph are represented as conditional probabilities. This tutorial will through examples give an informal introduction to theory and use of CPNs in connection with decision theory. The participants will obtain hands-on experience with the construction of a small CPN, including the acquisition of structure and conditional probabilities.

Tutorial presenters:
Finn V. Jensen
Dept. of Computer Science, Aalborg University,
Fredrik Bajers Vej 7,DK-9000 Aalborg Øst, Denmark.
Telephone: +4596358903,
email: fvj@cs.auc.dk

Steen Andreassen
Dept. of Medical Informatics and Image Analysis, Aalborg University,
Fredrik Bajers Vej 7D, DK-9000 Aalborg Øst, Denmark.
Telephone: +4596358812, Fax: +4598154008,
email: sa@miba.auc.dk


5) FOUNDATIONS OF PREFERENCE THEORY AND QUALITY OF LIFE ADJUSTMENT.

The methods of preference assessment and quality of life adjustment are widely applied in the medical decision making and cost-effectiveness literature. Yet, the theory and assumptions that underlie the use of these methods are poorly understood. The objectives of this short course are to provide experienced practioners with a quick and accessible introduction to the underpinnings of utility theory, with an emphasis on the relevance, power, and limitations of these assumptions in health and medical contexts. Topics to be covered will include: the theory of choice and preference; traditional models of individual decision making under uncertainty, including the von-Neumann - Morgenstern expected utility framework; the additional assumptions that support the use of multi-attribute utility functions and quality-adjusted life-years; and the difficulties encountered when the theory is extended beyond the individual to represent choice at the societal level.

Tutorial presenter:
Jospeh S. Pliskin, Ph.D.
Department of Industrial Engineering and Management and Department of Health Policy and Management,
Ben-Gurion University of the Negev,Beer-Sheva, Israel.
P.O. Box 653, Beer-Sheva 84105, Israel.
Telephone: 972-7-6472219,Fax: 972-7-6472958
email: jpliskin@bgumail.bgu.ac.il


6) HOW TO READ (AND MAYBE PERFORM) A SYSTEMATIC REVIEW (METAANALYSIS)

Physicians are committed to manage their patients according to the best available evidence. Systematic reviews are about asking the relevant questions; obtaining the published material (all of it); and extracting the evidence. In the tutorial we will address the following questions:

  1. Why do we need systematic reviews?
  2. How to put the questions?
  3. How to formulate a relevant protocol?
  4. How to collect the pertinent studies?
  5. How to evaluate the methodological soundness of the studies? Does it matter?
  6. How to obtain data from the studies and how to combine it?
  7. How to explore heterogeneity and why is it so important?
  8. How to check for biases?
  9. How to present results?
  10. Does metaanalysis work?
Tutorial presenters:

Karla Soares Weiser, Leonard Leibovici
Department of Medicine E, Beilinson Hospital,
Petah-tiqva 49100, Israel.
Telephone 972 3 9376501; fax 972 3 9376505;
email: leibovic@post.tau.ac.il

Invited Speakers

In addition to the above activities AIMDM´99 includes lectures from eminent speakers in the fields of Artificial Intelligence and decision making in medicine. Preliminary titles for these lectures are:

Accepted Papers (25 mins oral presentation)

Accepted Abstracts (15 mins oral presentation)

Accepted Abstracts (poster session)

Accepted Posters


[ Back ]