« back

Invited Talk

 

Health-e-Child CaseReasoner:

Similarity Search-based Clinical Knowledge Discovery and Decision Support with Neighborhood Graphs and Learning Discriminative Distance Functions

Alexey Tsymbal
Siemens AG, Corporate Technology
Erlangen, Germany

 

Abstract

In the context of the European funded project Health-e-Child, a Grid-based healthcare platform for European paediatrics is being developed. The basic philosophy behind the design of CaseReasoner, a CBR DSS system we are developing for Health-e-Child, is to provide a clinician with a flexible and interactive tool to enable operations such as data filtering and similarity search over a Grid of clinical centres, and also to facilitate the exploration of the resulting sets of clinical records. The aim is to let the clinician explore and compare the patients' records regardless of their geographical location, and to visualize their place in the distribution of the whole population of patients and of its semantic subgroups. For similarity search on distributed biomedical data, besides the canonical distance functions novel techniques for learning discriminative distance functions are also made available to the clinician. The use of distance learning techniques in combination with the patient similarity visualization modules of CaseReasoner contributes to making it a powerful tool for clinical knowledge discovery and decision support in various classification contexts; it helps to combine the power of strong learners with the transparency of case retrieval and nearest neighbor classification. In our talk the focus will be on presenting the two key technologies exploited in CaseReasoner; neighborhood graphs and learning discriminative distance functions. Examples will be given and our experience in the application of these techniques to various biomedical problem domains shared.

 

Vita

Alexey Tsymbal was born in Kharkiv, Ukraine, in 1975. He received the university diploma with distinction (in software engineering) from Kharkiv National Technical University of Radio-Electronics, Ukraine, in 1997, and the PhD degree in computer science from the University of Jyväskylä, Finland, in 2002. He is currently a Research Scientist at Siemens AG, Erlangen, Germany. He has more than 50 peer-reviewed publications in the areas of his research interests which include machine learning and data mining, knowledge-based techniques and systems, and applications of AI techniques to biomedical domains.
Dr Tsymbal is a member of Steering Committee for IEEE CBMS, the IEEE Symposium on Computer-Based Medical Systems, since 2005. He was General Co-chair of IEEE CBMS 2005, PC Co-Chair of CBMS 2008 and Special Tracks Chair of CBMS 2006 and 2007. Dr Tsymbal was a PC member at a number of leading conferences and workshops (including ICAIS 2009, eHealth 2009, ECML/PKDD 2008, ITAB/IS3BHE 2008, IJCAI 2007, BADE 2007, DMB 2007, ADMKD 2005-2007, ECAI 2006, PMKD 2005-2006) and served as a reviewer for a number of journals (including Artificial Intelligence Journal, IEEE Transactions on Knowledge and Data Engineering; IEEE Transactions on Neural Networks; IEEE Transactions on Pattern Analysis and Machine Intelligence; IEEE Transactions on Systems, Man, and Cybernetics; Information Fusion, and Medical Science Monitor). He acts also as Associate Editor for IEEE Transactions on Information Technology in Biomedicine.