Return-Path: <@csunb0.leeds.ac.uk:agc@scs.leeds.ac.uk> Received: from csunb0.leeds.ac.uk by paris.ics.uci.edu id aa04317; 14 Nov 97 10:50 PST Received: from csparc18.scs.leeds.ac.uk (csparc18.leeds.ac.uk [129.11.144.58]) by csunb0.leeds.ac.uk (8.6.12/8.6.12) with SMTP id SAA22862; Fri, 14 Nov 1997 18:44:33 GMT Date: Fri, 14 Nov 97 18:44:32 GMT Message-Id: <5660.9711141844@csparc18.scs.leeds.ac.uk> To: smyth@sifnos.ics.uci.edu Subject: aaai98 tutorial proposal From: Cohn A G Proposal to hold a Tutorial at AAAI98 on Qualitative Spatial Reasoning ======================================================================= Anthony G Cohn Professor of Automated Reasoning Division of AI, School of Computer Studies, University of Leeds, LEEDS, LS2 9JT, ENGLAND PHONE: +44 113 233 5482 (Secretary: 5437; Fax: 5468) agc@scs.leeds.ac.uk http://www.scs.leeds.ac.uk/spacenet/agc.html ======================================================================== Brief Description of the Tutorial ================================= While the field of Qualitative Reasoning is now well established it is only recently that there has been substantial investigation of calculi suitable for representing and reasoning about space in a qualitative way. This tutorial will survey the state of the art in qualitative spatial representation and reasoning techniques, including mechanisms for reasoning about spatial change. We will look at how these ideas may be applied in a wide variety of domains, from robotics and high level vision to the semantics of natural language expressions and of visual programming languages though to Geographical Information Systems. The tutorial will conclude with a look at open research issues. Why the Tutorial may be of interest to the AAAI98 audience. ========================================================== I believe a tutorial on Qualitative Spatial Reasoning at AAAI98 is timely for many reasons. Whereas temporal reasoning has now received extensive attention within various tutorial forums, there have been correspondingly very few presentations on (qualitative) spatial reasoning. In fact the tutorial at IJCAI95 by Hernandez and Mukerjee was almost certainly the first presentation at a major AI conference and while the amount of research in the subject is certainly increasing there have not been any tutorials since 1995 at a major AI conference, apart from my IJCAI97 tutorial on which this tutorial will build. Presenting this tutorial at AAAI98 would allow access to this material to a wider audience. Equally, the field is now maturing and the technology has advanced sufficiently so that the ideas can be used in various application domains, for example in Geographical Information Systems, high level vision, robotics, the semantics of spatial expression expressions, in specifying visual languages. There is now an active community working in the area, largely in Europe and to a certain extent the USA, but also in some other countries and there is a growing literature of substantive results. Brief Resume of the Proposer. ============================= Anthony G Cohn Professor of Automated Reasoning Division of AI, School of Computer Studies, University of Leeds, LEEDS, LS2 9JT, ENGLAND PHONE: +44 113 233 5482 (Secretary: 5437; Fax: 5468) agc@scs.leeds.ac.uk http://www.scs.leeds.ac.uk/spacenet/agc.html I hold a Personal Chair at the University of Leeds in Automated Reasoning and lead a research group one of whose current main activities is to research into qualitative spatial reasoning. I have been treasurer and Chairman of ECCAI, the European AI Coordinating Committee and have held the same positions in AISB, the UK AI society. I was the IJCAI95 workshop chair, ECAI94 programme chair and was on the advisory board for IJCAI93 so I believe I have a good appreciation of the nature of major conferences such as AAAI. I believe I am very well qualified to present an AAAI tutorial on QSR: I have been researching in QSR since the late 1980s and have published widely on the subject -- my web page lists all these publications (most are available as postscript). I have also held a number of grants to research into the subject, principally from the UK EPSRC, but I am also the Scientific Coordinator of SPACENET, a European Union funded HCM network on QSR which consists of 11 sites across Europe. Thus I have a good knowledge not only of the work at Leeds, but also at all these other sites and indeed others outside the network. Indeed in the SPACENET project we are writing a book on the state of the art in QSR and my tutorial will benefit from the extensive discussions we have held within our workshops -- and of course I gratefully acknowledge all the discussions I have had with my colleagues in the network. I have received various invitations to speak on QSR in various international forums: the 1995 Bolzano Spring School on Spatial Reasoning, invited talks at AISMC96 (Steyr, Austria), KI97 (Freiburg, Germany), a tutorial at FAPR96 (Bonn, Germany), co-workshop organiser/presenter at ESSLI95 (Barcelona, Spain). I have also been an invited participant at various closed workshops on spatial reasoning (I10 and I21 of the NSF-NCGIA; an GISDATA workshop in Vienna; an NSF workshop on spatial cognition; various IJCAI, ECAI and AAAI workshops). I have had extensive experience of lecturing at the Universities of Warwick and now Leeds since 1979 and gave a tutorial on Qualitative Reasoning jointly with Pat Hayes at ECAI86 and a lecture series at ACAI87 (Oslo, Norway) -- in addition to the various QSR tutorials mentioned above. I have written a survey article on QSR which appears in the proceedings of KI97 (and can be found on my web page) and also wrote a paper on the "Challenge of Qualitative Spatial Reasoning" which appeared in Computing Surveys in 1995 (this paper is also on my web page). Objective of the Tutorial and Intended Audience of the Tutorial =============================================================== The aim of the tutorial is to present the state of art in Qualitative Spatial Reasoning to the non specialist; this could either provide a suitable introduction for the intending research student, or for a practitioner in another research area or application domain who may be interested in exploiting QSR techniques. Some basic knowledge of logic would be helpful but not essential for understanding the tutorial. Proposed Structure of the Tutorial ================================== Introduction: Motivation for QSR: possible applications The interdisciplinary nature of QSR Spatial Ontlogy Some History Overview of the rest of the tutorial Aspects of Spatial Representations Topology Representation of topological information has been perhaps the most studied aspect of QSR, perhaps not unsurprisingly. The tutorial will cover these ideas in some detail. Distance and Size Although traditional QR scalar representations can be used for representing qualitative distances and sizes, there are various interesting pecularities to discuss when spatial representations are concerned. Orientation A number of orientation calculi have been developed; these will be presented and evaluated/compared. Shape Representing the shape of spatial entity qualitatively is hard, but there have been various interesting ideas, combining the calculi of topology, distance and orientation with specific notions of shape. These will be presented and compared. Uncertainty and Vagueness Many QSR calculi assume that although the calculus is qualitative, the underlying could in principle be specified precisely; this section of the tutorial investigates symbolic (non numerical) calculi which do not make this assumption: where there is an inherent vagueness in the knowledge of the spatial entities in question. Spatial Change This section of the tutorial will consider techniques for representing spatial change, particularly continuous change. Reasoning Techniques This section of the tutorial will cover various AI reasoning techniques such as constraints, various logical reasoning techniques and more specialised mechanisms such as composition tables. Metatheoretical results will also be covered here. Application Domains Time should allow a presentation of the application of QSR to some domains, hopefully illustrated by, for example, some video footage and/or some demos. Research Issues A brief guide to current and future research will be presented. Conclusions and Summary. <> I will make Powerpoint and postscript versions of some of my IJCAI97 tutorial available shortly at: http://www.scs.leeds.ac.uk/spacenet/ijcai97-tutorial.html ============================================= Tutorial Material ================= Copies of slides will be made available as will a short bibliography. Pointers will be given to online material including extensive online bibliographies. A/V requirements ================ I would like to present the tutorial using computer projection. Hopefully I can bring my own laptop to hook up to the computer projector or ohp panel, but ideally there would be a PC there (ideally which I could preload with my data via the internet, as the tutorial slides and demos are too big to put on a floppy or two). ======================================================================