SECOND CALL FOR PAPERS * Submission deadline: November 30, 2007 *
Special issue of _Information Fusion_ on Competing Fusion Methods - Real-world Performances ----- Submission procedures: http://perception.inrialpes.fr/people/Cuzzolin/files/IFJ-CFP-CFM-Final.pdf ----- This special issue will consider submissions that report on real-world experiences in which two or more distinct approaches to the same information fusion problem have been tested and compared. There are many competing methods for fusion, for example: Bayesian, statistical (such as frequentist, Kalman fliters, or Dempster-Shafer), heuristic (like templates, cluster analysis, or voting systems), or knowledge-based (such as fuzzy logic, rule-based, or other representations). Methods vary among themselves along several dimensions, such as generality, computational cost, how many and what specific type of assumptions are required, and whether those assumptions pertain only to the particular problem at hand. Further, it often happens that distinct theoretical approaches yield nearly identical conclusions, or equally plausible different conclusions. It can also happen that hybrid systems, which combine several approaches, might outperform systems based solely on any of the purebred approaches represented in the hybrid. An example of a hybrid could involve a hierarchical fuser in which some sensors combine by Bayes, other sensors by Dempster-Shafer, and these two sensor-groups are reconciled heuristically. We invite manuscripts that are neither previously published nor presented anywhere else in a more-or-less similar form. Manuscripts should clearly present: • A full description of the real-world information-fusion problem to which competing methods were applied • A description of the methods that were chosen for trial, including what specializations of the general methods were needed for the particular application (while avoiding tutorial-like presentations of wellknown methods) • The dimensions of performance studied (such as time requirements, reliability of inference, user understanding of results, and the like) • Additional information that contributed to the analysis (like theoretical considerations, or the experiences of others in the literature) • The methods and results of the testing • The conclusions drawn, including the choice of method, or combination of methods, ultimately fielded, with emphasis on possible new insights on the presented approaches gained through comparison. ----- _Information Fusion_ is an international journal on multi-sensor and multi-source information fusion, published by Elsevier. The Editor-in-Chief is Belur V. Dasarathy, Ph. D, FIEEE, [EMAIL PROTECTED], http://belur.no-ip.com The guest editors for this special issue are Fabio Cuzzolin, INRIA, [EMAIL PROTECTED] Paul Snow, [EMAIL PROTECTED] _______________________________________________ uai mailing list uai@ENGR.ORST.EDU https://secure.engr.oregonstate.edu/mailman/listinfo/uai
