From dcb@sequoia.wpi.edu Mon Mar 14 17:45:15 EST 1994 Article: 21075 of comp.ai Xref: glinda.oz.cs.cmu.edu comp.ai:21075 Path: honeydew.srv.cs.cmu.edu!fs7.ece.cmu.edu!europa.eng.gtefsd.com!MathWorks.Com!bigboote.WPI.EDU!cs!dcb From: dcb@sequoia.wpi.edu (David Brown) Newsgroups: comp.ai Subject: AID'94 Workshop on Machine Learning in Design Date: 12 Mar 1994 02:30:57 GMT Organization: Worcester Polytechnic Institute Lines: 153 Sender: dcb@sequoia (David Brown) Distribution: world Message-ID: <2lr9h1$km5@bigboote.WPI.EDU> NNTP-Posting-Host: sequoia.wpi.edu MACHINE LEARNING IN DESIGN ========================== workshop to be held prior to the THIRD INTERNATIONAL CONFERENCE ON ARTIFICIAL INTELLIGENCE IN DESIGN Swiss Federal Institute of Technology, Lausanne, Switzerland 15-18 August 1994 Design Learning vs Machine Learning ----------------------------------- ##### ##### ### ###### ###### ###### ### ### ### ### ### ### ### ### ### ### ### ### ### ### ### ## ### ### ### ###### ### ### ### ### ### #### ### ### ## ## #### ### ### ### ## ### ### ## ### ### ### ### ### ### ########### ## ## ## ### ### ### ### ########### ## ## ## ###### ======================================================================= Overview -------- The Machine Learning (ML) in Design workshop is part of the Artificial Intelligence in Design '94 conference. This workshop is intended to provide a forum for provocative discussion related to the application and evolution of machine learning techniques in design. To improve a product's quality or reduce the design cycle time designers require the right information, at the right time. Developments in computing technology present designers with ever increasing amounts of information and improved access. Therefore, the effective and efficient re-use of past design procedures and product knowledge is becoming more critical in contemporary design. Designers not only use information of specific experiences but also, by learning and understanding salient issues, they can abstract or generalise knowledge from those experiences. For example, they learn how to carry out design, what key decisions need to be made and when, what factors are crucial to decisions and what are the implications of those decisions, about the product itself, it's life expectancy and environment, it's manufacturability, the current state of technology, trends in the market place, and a lot more besides. Designers use their learned knowledge to create new designs (which may be novel or innovative) in an attempt to produce not only feasible but competitive products. Consequently, there is a growing realisation that for future computer-based design systems to be more effective they must continually evolve their state of knowledge to reflect new experiences and that they must use that knowledge in all aspects of design problem solving. By supporting the learning process it can be argued that computer based systems can become more effective tools which are better equipped to aid designers make well-informed decisions. Machine Learning, whose utility has been explored typically in chemistry, game-playing, image recognition, and many other fields, provides a basis to capitalise on the utility of inherent and explicit past design knowledge. However the application of machine learning in design is relatively immature and as such presents exciting and challenging issues to Machine Learning and Intelligent CAD researchers alike. ======================================================================= Purpose of the workshop ----------------------- The purpose of the workshop is to explore the issues and requirements of learning in design with a view of critically evaluating the current and required support from machine learning techniques. The objective is not only to identify key areas for future research but also to stimulate synergy in the Machine Learning in Design research community. ======================================================================= Workshop Format and Topics -------------------------- The workshop will run for half a day prior to the main AI in Design conference. Its format will depend upon submitted position papers but is likely to consist of a number of small working groups tackling identified issues and presenting their deliberations to the workshop participants for discussion. Topics may address issues such as: How can machine learning techniques be adapted or extended to support learning in design? What are the limitations of ML in Design ? Why does design present distinct challenges to ML research ? etc. ======================================================================= Guidelines for Position Papers ----------------------------- Position papers (in ascii or postscript form) of around three to four pages should be submitted by email to alex@cad.strath.ac.uk (Alex Duffy) no later than the 2nd June 1994. All accepted position papers will be distributed to the participants. ======================================================================= Number of Participants ---------------------- To stimulate lively debate and constructive discussions numbers will be restricted to around 20-30 participants. Admission, as determined by a selected panel, will be based upon the expediency of the submitted position papers and limited to attendees of the AI in Design conference. Each participant will be charged a fee of SFr 75 to cover costs of workshop notes, administration and refreshments. ======================================================================= Co-ordinators/Organisers ------------------------ Convenor: Dr Alex H B Duffy, University of Strathclyde, alex@cad.strath.ac.uk Co-Convenors: Prof David C Brown, Worcester Polytechnic Institute, dcb@willow.wpi.edu Prof Mary Lou Maher, University of Sydney, mary@archsci.arch.su.EDU.AU International Advisory Panel: Dr Tomasz Arciszewski (USA) Prof Ivan Bratko (Slovenia) Dr Ashok Goel (USA) Dr Yoram Reich (Israel) Prof Derek Sleeman (UK) ======================================================================= Further information ------------------- If you require further information then please contact: Dr Alex Duffy Postal Address: CAD Centre JANET: alex@cad.strath.ac.uk University of Strathclyde Phone:+44-41-552-4400 Ext. 3005 75 Montrose Street Fax :+44-41-552-3148 Glasgow G1 1XJ Scotland, UK. -------------------------------------------------------------------------- Article 21256 of comp.ai: Xref: glinda.oz.cs.cmu.edu comp.ai:21256 Newsgroups: comp.ai Path: honeydew.srv.cs.cmu.edu!nntp.club.cc.cmu.edu!news.mic.ucla.edu!library.ucla.edu!agate!howland.reston.ans.net!EU.net!sun4nl!news.nic.surfnet.nl!utciva.civ.utwente.nl!infnews.cs.utwente.nl!alberts From: alberts@cs.utwente.nl (Mert Alberts) Subject: AID94 Workshop: Semantic Basis for Sharing of Knowledge and Data in Message-ID: <1994Mar22.093538@cs.utwente.nl> Sender: usenet@cs.utwente.nl Nntp-Posting-Host: kopstoot.cs.utwente.nl Organization: Twente University, Dept. of Computer Science Date: Tue, 22 Mar 1994 08:35:38 GMT Lines: 142 Design Keywords: WORKSHOP ON SEMANTIC BASIS FOR SHARING OF KNOWLEDGE AND DATA IN DESIGN **************************************** to be held prior to the THIRD INTERNATIONAL CONFERENCE ON ARTIFICIAL INTELLIGENCE IN DESIGN Swiss Federal Institute of Technology, Lausanne, Switzerland 15-18 August 1994 ======================================================================= Summary -------- Design processes in general involve multiple participants from several disciplines and with different goals. As a consequence, a large number of different views on the required design product can be distinguished. Apart from syntactical differences due to the particular representational schemes employed for describing the design product, each of these views will bring along its own interpretation of the design object at hand. The problem of combining different representations, in terms of syntax and computational semantics, has already been extensively investigated in the domain of software engineering. Typical examples of established software engineering techniques for this purpose are parsing, filtering, and compiling. Less emphasis has been put on coping with different meanings or possible interpretations of the information (data and knowledge) represented. This distinction can be summarised as that between the computational aspects of information and the description of the concepts that this information refers to, respectively. Within AI, there seems to be a growing awareness of the importance of dealing with differences in the conceptual bases underlying the representation of information. For instance, several research groups are currently working on the development of so-called sharable ontologies: taxonomies of concepts, the `meaning' of which the users of the information have agreed upon. Most people seem to agree that communication about a problem requires some form of understanding of each others conceptualisation of that problem. However, there seems to be less agreement concerning the means for establishing this understanding. ======================================================================= Goals and Scope of the Workshop ------------------------------- In the workshop, we would like to inventory existing ontologies and conceptual models for design in the context of sharability. Typical questions would be: what are important sources for a shared conceptual basis (e.g. design standards, social conventions, and legacy systems), and: how do we guarantee that conceptualisations can be shared between different design agents. The goal of the workshop is to improve the semantic basis required for the development of software systems that allow for communication and sharing among engineering knowledge and data bases used by different members of a design team. Expected outcomes include an inventory of existing approaches to and recommendations for obtaining such a semantic basis. ======================================================================= Workshop Format --------------- The workshop will run over half a day prior to the main AI in Design conference. Accepted papers will be distributed prior to the workshop with a set of discussion topics. A selection will be made among the accepted papers, the authors of which will be asked to present a (very) concise summary of their views, concentrating on the discussion topics. These presentations are intended as ``stepping stones'' for extensive discussion, rather than at given a detailed account of the specific approach taken. ======================================================================= Guidelines for Submissions -------------------------- Research-summary or position papers (in ascii or postscript, 8 pages max.) should be submitted by email to Mert Alberts, alberts@cs.utwente.nl, no later than the 2nd June 1994. Submitted papers must clearly state the suggested approach to the problem of interfacing different conceptual models of a design product, and focus around one of the following topics: - conceptual models of design - ontologies for design knowledge - combining different ontologies - multiple views on design products - sharable representations of design objects and design knowledge ======================================================================= Number of Participants ---------------------- There will be 20-25 participants. Admission, as determined by a selected panel, will be based upon the expediency of the submitted position papers and limited to attendees of the AI in Design conference. Each participant will be charged a fee of SFr 75 to cover costs of workshop notes, administration and refreshments. ======================================================================= Organisation -------------------- Organising Committee: Mert Alberts University of Twente The Netherlands e-mail: alberts@cs.utwente.nl Tom Gruber Stanford University USA e-mail: Gruber@KSL.Stanford.edu Mary-Lou Maher University of Sydney Australia e-mail: mary@archsci.arch.su.EDU.AU International Advisory Committee: Bernd Bachmann, DFKI, Germany James Bowen, National University at Cork, Ireland Cherif Branki, Glasgow Caledonian University, UK Susan Finger, Carnegy Mellon University, USA Mike Rosenman, University of Sydney, Australia Duvvuru Sriram, MIT, USA Chris Tong, Rutgers University, USA Nel Wognum, University of Twente, The Netherlands -- L.K. Alberts Tel: +31 53 89 4291 University of Twente Department of Computer Science Fax: +31 53 339605 P.O. Box 217 7500 AE Enschede The Netherlands E-mail: alberts@cs.utwente.nl Article 21661 of comp.ai: Xref: glinda.oz.cs.cmu.edu comp.ai:21661 sci.cognitive:3456 Path: honeydew.srv.cs.cmu.edu!nntp.club.cc.cmu.edu!miner.usbm.gov!rsg1.er.usgs.gov!darwin.sura.net!howland.reston.ans.net!gatech!swrinde!sgiblab!sgigate.sgi.com!olivea!news.hal.COM!decwrl!morrow.stanford.edu!hpp.Stanford.EDU!narayan From: narayan@hpp.Stanford.EDU (Hari Narayanan) Newsgroups: comp.ai,sci.cognitive Subject: CFP: AID'94 Workshop on Reasoning with Shapes in Design Date: 15 Apr 1994 00:29:04 GMT Organization: Center for Advanced Medical Informatics at Stanford Lines: 130 Distribution: world Message-ID: <2okn4g$gi0@morrow.stanford.edu> NNTP-Posting-Host: hpp.stanford.edu ============================================================== CALL FOR WORKSHOP PARTICIPATION Reasoning with Shapes in Design =============================== Sunday, August 14th, 1994 14:00--18:00 hs Lausanne, Switzerland Held in conjunction with AID'94 (Artificial Intelligence in Design '94) Description of Workshop ----------------------- 'Shape' -- models of geometry -- plays an important role in the development of design ideas. This workshop will provide a forum for researchers in artificial intelligence in design to discuss reasoning with shape in design and the representations and computations that support such reasoning. The workshop organisers especially invite attention to the following issues: * designers' understanding of shape information * the role of shape information in different models of design (examples of design models include search, constraints, prototypes, optimisation, and genetic processes) * models of shapes * representations of shapes (emerging representational schemes such as connectionist and diagrammatic representations as well as traditional symbolic representations, including mathematical ones, are of interest) * reasoning with shapes * transformations between alternative representations of shapes * computing with shapes * design applications involving shape Format of the Workshop ---------------------- The workshop will have a round-table format. Up to four 20-minute presentations based on extended abstracts submitted by participants will be given. Its duration will be 4 hours with a half hour break at the midpoint. The presentations are intended to act as catalysts for discussion which can and should go beyond the presentations themselves. Attendance ---------- Attendance at the workshop will be restricted to persons whose extended abstracts are accepted by the advisory committee. To help ensure a focussed and collegial atmosphere, the number of attendees will be limited to 20-30 people. Participants must register for both the workshop and the general conference. The workshop fee will be SFr 75. This is to cover administration, workshop notes and coffee break. Any questions regarding the technical content of the workshop should be directed to either of the coordinators. Submissions Requirements ------------------------ Submit an extended abstract of 2-5 pages. Authors whose abstracts are selected by the advisory committee for presentation will be requested to prepare 20-minute presentations of their work. Send submissions in plain ascii or Latex source format to jose@arch.su.edu.au, or mail 4 copies of hardcopy submissions to Jose Damski at the address below. Electronic submissions are preferred. Please do not send floppy disks or tapes. Important Dates --------------- 2 June 1994 Deadline for submissions. 4 July 1994 Notification of acceptance. 15 July 1994 Deadline for camera-ready copies. Publication ------------ The preprints of the workshop will be published, and distributed to the participants at the workshop. Authors of some abstracts may subsequently be invited to submit a full paper to a special issue of an archival journal related to the topic of reasoning with shapes. Workshop Committee ------------------ Christopher Carlson Research Institute for Symbolic Computation Johannes Kepler University ccarlson@risc.uni-linz.ac.at Jose Damski (coordinator) Key Centre of Design Computing Dept of Architectural and Design Science University of Sydney - Sydney - 2006 - NSW - Australia fax: +61-2-692 3031 jose@arch.su.edu.au Ernest Edmonds LUTCHI Research Centre Department of Computer Studies Loughborough University of Technology e.a.edmonds@lut.ac.uk Hari Narayanan Knowledge Systems Laboratory Stanford University narayan@hpp.stanford.edu George Stiny University of California - Los Angeles STINY@pace1.gsaup.ucla.edu Robert Woodbury (coordinator) Department of Architecture University of Adelaide rw@arch.adelaide.edu.au ============================================================== Article 22062 of comp.ai: Xref: glinda.oz.cs.cmu.edu comp.ai:22062 Newsgroups: comp.ai Path: honeydew.srv.cs.cmu.edu!nntp.club.cc.cmu.edu!newsfeed.pitt.edu!gatech!howland.reston.ans.net!pipex!warwick!bsmail!irix!enknb From: enknb@irix.bris.ac.uk (Ken Brown) Subject: cfp: Role of Theory in AI in Design Message-ID: Sender: usenet@info.bris.ac.uk (Usenet news owner) Nntp-Posting-Host: irix.bris.ac.uk Reply-To: Ken.Brown@bristol.ac.uk Organization: University of Bristol, England X-Newsreader: TIN [version 1.2 PL1] Date: Wed, 11 May 1994 13:57:20 GMT Lines: 119 ----------------------------------------------------------------------------- Workshop Call: ************* The Nature and Role of Theories of Design in AI in Design Research ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Workshop to be held as part of the Third International Conference on Artificial Intelligence in Design Swiss Federal Institute of Technology, Lausanne, Switzerland 15-18 August 1994 Background and Motivation: ~~~~~~~~~~~~~~~~~~~~~~~~~ The work reported in the AI in Design literature has so far shown little concern for the nature, development, or role of theories of design in AI in Design research. A notable exception to this is a paper by John Dixon, [Dixon 1987], in which he presents the view that theories of design should look like theories in the physical sciences. An implication of such a view is, that theories of design should be `hard' and testable, like those of physics, chemistry, etc. A further implication is that such theories can be used to make precise predictions that can be tested by a single well designed and executed experiment--as Newton's laws of motion can be, for example. Though Dixon presents clear arguments for his position, there seems to be little evidence (yet, at least) that we are seeing the development of this kind of theory of design in AI in Design research, nor is there any reporting of the outcome of decisive experiments that attempt to test any theories of design of this kind. A different view on the nature of design theory in AI in Design research is taken by those people who class the design process as a behavioural phenomenon, not as a physical one. In so doing, they advocate that theories of design should look more like those of the behavioural sciences. These are typically `soft' theories which do not support precise value predictions, but rather, attempt to identify the essential components of behavioural phenomena, the relationships between them, and important qualitative general properties. The development and testing of these kinds of theories typically involves large numbers of experiments in which statistical variation and uncontrolled variables have a significant effect, thus making the interpretation of the results from one experiment difficult and potentially hazardous. However, while the AI in Design literature does contain some reporting of behaviour studies of human designers, it does not report on any work that can be understood as forming part of an on going attempt to develop and test this kind of theory of design. Why then is there apparently so little concern for the nature, development, and role of theories of design in the work reported in AI in Design? Is it because the field has no strong need for theories of design? Is it because what theory it does need it gets from elsewhere? Or, is it still too early to expect it to be able to effectively support this kind of work? The answer to all these questions would seem to be no, which leaves the issue of the nature and role of theories of design in AI in Design research in need of some urgent consideration and discussion. References: [Dixon, 1987] John Dixon, 1987. On research methodology towards a scientific theory of engineering design, AI-EDAM, vol 1, no 3, pp 145--157. Aims and Objectives of the Workshop: ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ The main aims of this workshop are to survey and review the current status of theories of design in AI in Design research and work devoted specifically to experimental testing of design theories. The objective is to try to bring together what work there is on this subject and to establish a basis for further fruitful interaction and discussion of the issues involved by the people involved. Format: ~~~~~~ The workshop will be organised around a number of short presentations and discussions of contributions selected from those submitted. With the selection being based on an attempt to have all the major issues and aspects raised in the contributions represented. Guidelines for Submissions: ~~~~~~~~~~~~~~~~~~~~~~~~~~ People wishing to attend the workshop should submit a clear and concise position paper (up to five pages) which: describes work aimed at developing a theory of design in the context of AI in Design; or constructively comments on some existing and published theoretical developments; or describes work aimed at experimentally testing any theory of design or some significant part or aspect of one (not necessarily from AI in Design); or presents an argument for what kind of theory (or theories) of design is (are) needed in AI in Design, and why; or presents an argument for why no theoretical work is required in AI in Design, and why; or any other kind of contribution which the author(s) feel makes a constructive contribution to the subject of the workshop. Submitted contributions will be reviewed by the organising committe and selected on the basis of clarity, conciseness, and relevance. They should be sent as ascii of Latex source by email to: Tim Smithers at ccpsmsmt@si.ehu.es to arrive by the 2 June, 1994, to whom all enquires and questions may be also sent. Workshop Organising Committee: ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Subrata Dasgupta, University of Southwestern Louisiana, USA Susan Finger, Carnegie Mellon University, USA Ken MacCallum, Strathclyde University, Scotland Tim Smithers, Universidad del Pais Vasco, The Basque Country Wade Troxell, Colorado State University, USA