Abstracts:

Patrick Saint-Dizier
IRIT-CNRS, 118, route de Narbonne 31062 Toulouse Cedex France stdizier@irit.fr

In this paper, we firstly outline some elements related to sense variation and to sense delimitation within the perspective of the Generative Lexicon. We then develop the case of adjectival modification and several forms of selective binding and metonymies for verbs and show that, in some cases, the Qualia structure can be combined with or replaced by a small number of rules, which seem to capture more adequately the relationships between the predicator and one of its arguments. We focus on the Telic role of the Qualia structure which seems to be the most productive role to model sense variations. (pp. 409-422)

Keywords: Lexical semantics, meaning variations

George Stephanopoulos
Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge, MA 02139, U.S.A. geosteph@mit.edu

A computer-aided environment for the synthesis of processing schemes for the production of pharmaceuticals is presented. Human developers are guided through the innovative development process through a hierarchy of decision-levels following the principle of hierarchical decomposition. Progress of the development is organized through a knowledge-based model of process chemicals and their transformations involving quantitative and qualitative knowledge. Process chemists can directly implement their processing ideas by using a ``natural" language used for the description of lab recipes. The computer-aided environment maintains consistency with first principles, offers assistance for crucial design decisions and solves autonomously specific sub-problems. The presented methodology demonstrates an example for the successful man-computer interaction in the innovative task of process synthesis. (pp. 423-434)

Keywords: Batch design, Natural Language, Task-Oriented Design

Paul Tarau
Department of Computing Science, University of Moncton, Moncton,NB, Canada E1A 3E9 tarau@info.umoncton.ca

In this paper\footnotemark we examine some knowledge base uses of a recently developed logic grammar formalism, Assumption Grammars, particularly suitable for hypothetical reasoning. They are based on intuitionistic and linear implications scoped over the current continuation, which allows us to follow given branches of the computation under hypotheses that disappear when and if backtracking takes place. In previous research on using Assumption Grammars [TDF96] for processing natural language [DTL97], we showed that they allow abstracting time and consumer/producer relationships in complex natural language processing (relatives, anaphora), therefore resulting in more readable programs; and that they offer the flexibility of switching between data-driven or goal-driven reasoning, at no overhead in terms of either syntax or implementation. In this paper we argue that Assumption Grammars are also useful for knowledge-based systems, and that some of the techniques developed for natural language processing can naturally and usefully be transferred to knowledge based systems applications. Surprisingly, the technique may also be extended to coordination languages (Linda) with similar results. (pp. 435-444)

Keywords: Logic grammars, natural language processing, intuitionistic and linear implication

R. Sahandi, D.S.G. Vine and J. Longster
Bournemouth University, Fern Barrow, Poole, BH12 5BB, UK, Tel.: +44 1202 595258, Fax: +44 1202 595314 rsahandi@bournemouth.ac.uk

The development of interactive multimedia systems, coupled with the advances in computer technology and high-speed communication systems, has made it possible for information to be presented to users more effectively and efficiently. Man-machine communication can be enhanced via the use of synthesised speech and computer animation in multimedia systems. Whilst synthetic speech is potentially a more natural communication medium, it can be improved by the addition of an animated human face synchronised with the synthetic speech. This facial display provides a number of visual cues relating to what the speaker is saying and the speaker's emotional state. This increases intelligibility if the synthetic speech is degraded with noise, and allows knowledge transfer for the hearing-impaired, through lip-reading. This paper provides an overview of existing speech synthesis and facial animation techniques, and discusses the limitations of each. The paper concludes with a description of a visual speech synthesis system developed at Bournemouth University, and a discussion of the audio-visual synchronisation issues. (pp.445-450)

Keywords: Synthetic speech, Facial animation

Vladimir A. Fomichov
Faculty of Applied Mathematics, Moscow State Institute of Electronics and Mathematics (Technical University), 109028 Moscow, Russia and Department of Discrete Mathematics, Faculty of Mechanics and Mathematics Lomonosov Moscow State University, 119899 Moscow, Russia Tel: 007-095-930-9897; Fax: 007-095-939-2090 E-mail: vladfom@yahoo.com, vaf@nw.math.msu.su

For the first time, a comprehensive and flexible mathematical framework for constructing agent communication languages (ACLs) is suggested. That framework is the theory of restricted K-calcules and K-languages, or the RKCL-theory, published in Informatica, 1996, No. 1. The essence of the new framework is as follows. 10 operations on formal representations of conceptual structures are suggested. The contents of arbitrary messages and arbitrary communicative acts may be represented by the strings built out of primitive conceptual items by means of these 10 operations. The RKCL-theory possesses the expressive possibilities exceeding the expressive possibilities of the ACL published in October 1997 in Geneva by the international Federation of Intelligent Physical Agents (FIPA). The main advantages are, in particular, the possibilities to represent arbitrarily complicated goals (actions) and structured meanings of arbitrary discourses in natural language. The new framework enables us both to define syntax of ACLs and to formalize their semantics. The paper suggests also a little modified variant of the RKCL-theory without changing the expressive power of that theory. (pp. 451-464)

Keywords: multi-agent system, communicative act, agent communication language, natural language, syntax, semantics, electronic commerce, Internet, FIPA ACL, integral formal semantics, restricted K-calculus, restricted standard K-language, ontology, semantic representation, structured meaning, discourse, knowledge representation, natural language processing, conceptual processing

M. Magnanelli, A. Erni and M. Norrie
Institute for Information Systems, CH-8092 Zurich, Switzerland {magnanel, erni, norrie}@inf.ethz.ch

We describe an Internet agent which continuously gathers information from Web documents in order to maintain a local database and ensure its currency. As a specific application, we detail an agent which maintains a database with information about academic contaCTS, their projects and publications. Agent operation is driven by a combination of an extraction profile which specifie what and how information is to be extracted from Web documents and local database which specifies the particular contacts of interest. The agent detects both new and updated information and, when the confidence level is above a user-specified threshold, automatically updates the local database accordingly. (pp. 465-474)

Keywords:agents, information extraction

Silvano Mussi
CILEA - Via Sanzio 4, I - 20090 Segrate (MI), Italy mussi@icil64.cilea.it

In this paper, we propose an approach to the design of multi-agent systems based on an original model of the mental activity of single agents. In particular, we introduce the concept of active mental entity, as a new way of representing mental attitudes such as intentions and persuasions. The internal architecture of each agent is thus understood as a distributed system whose reasoning activity is determined by the interactions among active mental entities. Then, this architecture is extended in order to enable the agents to operate in a multi-agent context. A detailed description of the structure and operation of an agent and of a multi-agent system is thus provided. The implementation of the proposed paradigm is then illustrated and some performance examples are presented. (pp. 475-484)

Keywords:distributed artificial intelligence, multi-agent systems, mental attitudes

Marek Machura
Dept. of Computing, Bournemouth University, Talbot Campus, Poole, Dorset BH12 5BB, UK

This paper describes an implementation of the Finite Difference Time Domain (FDTD) method using MPI (Message Passing Interface) library to the sample electromagnetic problem: calculating the input impedence of a microstrip antenna. Its objective is to demonstrate to engineers in disciplines other than computing, how easy it is to use contemporary software tools, such as MPI, to speed up engineering computations. (pp. 485-490)

Keywords:Distributed Computing, MPI, Electromagnetic Field Calculations, FDTD

Haun-Chao Keh and Jen-Chih Lin
Department of Computer Science and Information Engineering, Tamkang University, Tamsui, Taipei, Taiwan 251, R.O.C. g5190032@tkgis.tku.edu.tw

We consider a new supercube architecture, a new interconnection network derived from the hypercube. The supercube retains the connectivity and diameter properties of the corresponding hypercube. The embedding of one interconnection network into another is a very important issue in the design and analysis of parallel algorithms. In this paper, the problem of embedding and reconfiguring complete binary tree structures is considered in a supercube with faulty nodes. We also propose a new method for embedding and reconfiguring complete binary trees in a faulty supercube. Furthermore, the results enable us to obtain a good method for embedding complete binary tree structures into a faulty supercube with n-expansion. The result enables us to obtain O(n2-m2) faults which can be tolerated, where (n-1) is the dimension of a supercube and (m-1) is the height of a complete binary tree. (pp.491-498)

Keywords:parallel algorithm, fault-tolerance, supercube, embed

Jie Wu
Department of Computer Science and Engineering, Florida Atlantic University, Boca Raton, FL 33431, USA

A new class of interconnection networks, the hypernetworks, havebeen proposedrecently. Hypernetworks are characterized by hypergraphs. Compared with point-to-pointnetworks, they allow for increased resource-sharing and communication bandwidth utilization, and they are especially suitable for optical interconnects. One way to derive a hypernetwork is by finding the dual of a point-to-point network. Hypercube Qn, where n is the dimension, is a very popular point-to-point network. In this article, we consider using the dual Q*n of hypercube of Qn as an interconnection network. We investigate the properties of Q*n, and present a set of fundamental data communication algorithms for Qn*. Our results indicate that hypernetwork Qn* is a useful and promising interconnection structure for high-performance parallel and distributed computing systems. (pp. 499-508)

Keywords: hypercube, hypergraph, hypernetwork, interconnection network, optical bus, optical interconnection, parallel and distributed computing