Scope of the Conference: The conference seeks original research and application papers in ALL areas of Artificial Intelligence. Possible topics can be found here. Submission Guidelines: Conference submission is electronic in postscript format. (Please, do not submit PDF.) Submitted papers must not exceed 15 pages and should conform to the Springer LNCS style. Authors are strongly encouraged to use LaTeX2e and the Springer llncs class files. If you have difficulties producing a Springer LNCS style document or you cannot produce postscript output, please contact ki2002@rwth-aachen.de. To electronically submit your paper and abstract, please follow the instructions on our submission web-form (deadline elapsed). Important Dates for Authors: 12 March - 09 April 2002 Paper submission page accessible 09 April 2002 Deadline for paper submissions 27 May 2002 Notification of acceptance 25 June 2002 Final paper due 16-20 September 2002 KI2002 Reviewing Criteria: Papers will be evaluated according to the following criteria. Please keep these criteria in mind when preparing your submission. 1. Papers on Foundations Papers on foundations should present significant and original results that are relevant to AI. The main criterion for inclusion of a paper in the proceedings is whether sufficiently many readers can be expected to profit from reading the paper for their future AI research. In addition to this general criterion, the reviewers will be asked to evaluate the papers according to the following specific criteria: RELEVANCE: Is this paper relevant to AI researchers or does it address issues not relevant for AI? Would it perhaps better be presented at another conference? Is the paper accessible to a general AI audience, or is it only comprehensible for a small segment of the community? SIGNIFICANCE: How important is the work reported in the paper? Does it attack an important and difficult problem or a peripheral and simple problem? The problem should be interesting and natural, and not just be chosen by the authors because it can be attacked by their methods. Does the approach used to solve the problem advance the state of the art or is it just one more application of a well-known method? ORIGINALITY: Has this or similar work been previously reported? Are the problems and/or approaches completely new? Is it a novel combination of familiar techniques? Does the paper introduce new approaches and methods, or is it reinventing the wheel using new terminology? The paper should clarify the relationship to previous work. Even if the authors think that previous work in this direction is irrelevant for their work, they should justify their opinion. QUALITY: Is the paper technically sound? Not only the results but also the proofs must be correct. Does the paper contain enough details to check for correctness? If not, do the authors cite an accessible technical report containing full proofs? In addition to theoretical analysis, empirical evaluations (e.g., of the claim that a certain algorithm behaves good in practice) are also relevant here. PRESENTATION: Is the paper well organized and well written? The information of the paper should be available to the reader with a minimum of effort. Does the paper cite and discuss previous work? Does it motivate the problem and the approach chosen to solve it? Does it make clear what has been achieved with this paper and what is still to be done? Are definitions crisp and to the point, or does the paper introduce notation after notation without really working with the introduced terminology? Even technical papers on a narrow topic should be written such that non-experts can comprehend the main contribution of the paper and the methods employed. The paper shouldn't just be a litany of deep but obscure theorems. 2. Papers on Applications Papers on applications should present significant and original applications of AI methods and tools. The main criterion for inclusion of a paper in the proceedings is whether sufficiently many readers can be expected to profit from reading the paper for their own application. Thus, it is not enough if all one can learn from the paper is that the authors have an application. Someone with a similar application should be expected to have a considerable advantage after having read the paper. In addition to this general criterion, the reviewers will be asked to evaluate the papers according to the following specific criteria: RELEVANCE: Is the application built using AI methods, or is it more or less a standard computer science application? Is the use of AI methods justified? SIGNIFICANCE: How important is the application reported in the paper? Does it involve important and difficult problems or peripheral and simple problems? The application should come from the "real world", and not just be invented by the authors because it can be solved using their methods. If the authors consider a "toy application", what is the purpose? What new insights (e.g., on the behavior of an algorithm) can be gained from considering the toy application? Can one expect the results to generalize to "real world applications"? Does the approach used to build the application advance the state of the art or is it just one more application of a well-known method? ORIGINALITY: Has this or similar work been previously reported? It is not necessary for the paper to develop new AI techniques, but it should at least apply them in a novel way or shed a new light on their applicability in a certain domain. The paper should clarify the relationship to previous work. Even if the authors think that previous work in this direction is irrelevant for their work, they should justify their opinion. QUALITY: Is the approach chosen by the authors sound? It is not enough to say that a certain technique has been applied, but it must also be evaluated whether and why this technique worked in the application. The chosen approach should be described in enough detail to allow such an evaluation. Are there empirical evaluations that support the claims made in the paper? PRESENTATION: Is the paper well organized and well written? The information of the paper should be available to the reader with a minimum of effort. Does the paper cite and discuss previous work? Does it motivate the application and the approach chosen to solve it? Does it make clear what has been achieved with this paper and what is still to be done?