Return-Path: Received: from CS.CMU.EDU by A.GP.CS.CMU.EDU id aa01034; 8 Feb 94 0:40:17 EST Received: from Csli.Stanford.EDU by CS.CMU.EDU id aa06859; 8 Feb 94 0:39:52 EST Received: from localhost.Stanford.EDU by CSLI.Stanford.EDU (4.1/25-CSLI-eef) id AA27298; Mon, 7 Feb 94 20:31:23 PST Message-Id: <9402080431.AA27298@CSLI.Stanford.EDU> From: vieu@irit.irit.fr (Laure Vieu) (by way of yarowsky@unagi.cis.upenn.edu (David Yarowsky)) Subject: ECAI Workshop on Parts and Wholes To: empiricists@CSLI.Stanford.EDU Date: Mon, 07 Feb 1994 20:31:20 -0800 Sender: roscheis@CSLI.Stanford.EDU CALL FOR WORKSHOP PARTICIPATION PARTS AND WHOLES: CONCEPTUAL PART-WHOLE RELATIONS AND FORMAL MEREOLOGY Monday, August 8, 1994 Amsterdam, The Netherlands Held in conjunction with ECAI-94 (11th European Conference on Artificial Intelligence) Currently, there are two main approaches to the study of "parts" and their relations. The conceptual (cognitive) approach looks at the variety of part-whole relations and their role in language processing, perception, and action planning; the philosophical/logical approach, on the other hand, looks at formal theories of parts, wholes and related concepts in the framework of formal ontology. There are important differences between the two views. Philosophical systems tend to focus on a single "part-*of*" relation used for modeling ontological domains like time, space, or pluralities; conceptual approaches tend to assume a whole family of different "part-*whole*" relations for a variety of entities and tasks. Classical logical theories such as Lesniewski's or Goodman's privileged extensional aspects of the part-wholerelation, while for conceptual approaches and intensional formal mereology the old proverb holds that a whole is more than its parts. While disciplines such as linguistics, philosophy and psychology have contributed significantly to the research in this field, their impact on artificial intelligence is extremely limited, although AI could represent the ideal workbench for a unification of approaches dominant in different fields. Knowledge about parts is of great importance for a wide variety of AI domains, like vision, qualitative and naive physics, robotics, and natural language processing. For example, the structure of an object can be used for visual recognition, for reasoning about the functionality of the whole, or for planning its assembly. The goal of this workshop is to bring together researchers from these various disciplines in order to explore (i) the benefits and limits of formal mereology in modeling commonsense part-whole relations; (ii) the import for knowledge representation formalisms of the two current approaches to the study of "parts" and their relations; (iii) the possibility of a unified theory of parts and wholes. The workshop topics will include the following: - Classical extensional mereology: uses, extensions and adaptations; comparison with alternatives to mereology: set theory and lattices. - Intensional mereology: essence, dependence, and integrity; individual properties as parts. - Conceptual distinctions among wholes: masses, collections, complexes; natural entities and artifacts; sums and scattered individuals. - Parts and structure: physical connection, spatial, temporal, functional and other constraints among parts; Gestalt theories and perceptual parts; granularity issues. - Parts, space and time: relationships between mereology, topology, geometry; boundaries and surfaces; relationships between parts of physical objects (continuants) and parts of events (occurrents). - Parts and natural language: parts, part-names and possessive constructions; plurals and mass terms. - Reasoning about parts: transitivity, upper and downward inheritance of properties. - Dealing with parts within existing KR formalisms: distinguishing parts from other attributes, computational issues of reasoning about parts. Two possible kinds of contributions are solicited from interested participants: (a) regular papers of 10 pages max, presenting on-going research; (b) position papers of 3 pages max, motivating the interest in the field and explaining particular points of view. A limited number of regular papers will be chosen for an oral presentation at the workshop, while suitable space will be devoted to discussions based on contributions from participants (rejected regular papers are automatically treated as position papers). Participation will be limited to around 35 people. Preference will be given in the workshop schedule to contributions underlining the impact of mereological issues on AI practice, especially on: knowledge representation, natural language processing, qualitative and naive physics, spatial and temporal reasoning, vision, and robotics. Submission of papers, regular and position, to any member of the workshop organizing committee is due by April 15 1994. Hard copy (4 copies) and electronic submissions (either PostScript, LaTex or MacWord converted in BinHex format) are equally acceptable, with a strong preference for the latter. All submissions should include an exact address and an e-mail address. TIMETABLE: Paper submission deadline: April 15, 1994 Notification: May 20, 1994 Final version due: June 6, 1994 Workshop: August 8, 1994 IMPORTANT NOTICE: Participants will be requested to register for the main ECAI conference. Organizing committee: Nicola Guarino LADSEB-CNR Corso Stati Uniti 4, I-35020 Padova tel: +39 49 8295751, fax: +39 49 8295778 email: guarino@ladseb.pd.cnr.it Simone Pribbenow University of Hamburg, Computer Science Department, Bodenstedtstr. 16, D-22765 Hamburg tel: +49 40 4123-6111, fax: +49 40 4123-6159 email: pribbeno@informatik.uni-hamburg.de Laure Vieu Institut de Recherche en Informatique de Toulouse, UPS, 118 route de Narbonne, F-31326 Toulouse tel: +33 61556091, fax: +33 61558325 email: vieu@irit.fr