The purpose of this paper is to explore some semantic problems related to the use of linguistic ontologies in information systems, and to suggest some organizing principles aimed t o solve such problems. The taxonomic structure of current ontologies is unfortunately quite complicated and hard to understand, especially for what concerns the upper levels. I will focus here on the problem of ISA overloading, which I believe is the main responsible of these difficulties. To this purpose, I will carefully analyze the ontological nature of the categories used in current upper-level structures, considering the necessity of splitting them according to more subtle distinctions or the opportunity of excluding them because of their limited organizational role.

Abstract. (Semi-)automatic mapping — also called (semi-)automatic alignment — of ontologies is a core task to achieve interoperability when two agents or services use different ontologies. In the existing literature, the focus has so far been on improving the quality of mapping results. We here consider QOM, Quick Ontology Mapping, as a way to trade off between effectiveness (i.e. quality) and efficiency of the mapping generation algorithms. We show that QOM has lower run-time complexity than existing prominent approaches. Then, we show in experiments that this theoretical investigation translates into practical benefits. While QOM gives up some of the possibilities for producing high-quality results in favor of efficiency, our experiments show that this loss of quality is marginal. 1

. Adaptation in open, multi-agent information gathering systems is important for several reasons. These reasons include the inability to accurately predict future problem-solving workloads, future changes in existing information requests, future failures and additions of agents and data supply resources, and other future task environment characteristic changes that require system reorganization. We have developed a multi-agent distributed system infrastructure, Retsina (REusable Task Structure-based Intelligent Network Agents) that handles adaptation in an open Internet environment. Adaptation occurs both at the individual agent level as well as at the overall agent organization level. The Retsina system has three types of agents. Interface agents interact with the user receiving user specifications and delivering results. They acquire, model, and utilize user preferences to guide system coordination in support of the user's tasks. Task agents help users perform tasks by formulating p...

To facilitate the rapid development and open system interoperability of autonomous agents we need to carefully specify and effectively implement various classes of agent behaviors. Our current focus is on the behaviors and underlying architecture of WWW-based autonomous software agents that collect and supply information to humans and other computational agents. This paper discusses a set of architectural building blocks that support the specification of behaviors for these information agents in a way that allows periodic actions, interleaving of planning and execution, and the concurrent activation of multiple behaviors with asynchronous components. We present an initial set of information agent behaviors, including responding to repetitive queries, monitoring information sources, advertising capabilities, and self cloning. We have implemented and tested these behaviors on the WWW in the context of WAR-REN, an open multi-agent organization for financial portfolio management.

We describe in this paper the ML system, ASIUM, which learns subcategorization frames of verbs and ontologies from syntactic parsing of technical texts in natural language. The restrictions of selection in the subcategorization frames are filled by the concepts of the ontology. Applications requiring subcategorization frames and ontologies are crucial and numerous. The most direct applications are semantic checking of texts and syntactic parsing improvement but also text generation and translation. The inputs of ASIUM result from syntactic parsing of texts, they are subcategorization examples and basic clusters formed by head words that occur with the same verb after the same preposition (or with the same syntactical role). ASIUM successively aggregates the clusters to form new concepts in the form of a generality graph that represents the ontology of the domain. Subcategorization frames are learned in parallel, so that as concepts are formed, they fill restrictions of selection in th...

Ontology mapping is important when working with more than one ontology. Typically similarity considerations are the basis for this. In this paper an approach to integrate various similarity methods is presented. In brief, we determine similarity through rules which have been encoded by ontology experts.

Abstract not found

Abstract not found

The task of information extraction can be seen as a problem of semantic matching between a user-defined template and a piece of information written in natural language. To this purpose, the ontological assumptions of the template need to be suitably specified, and compared with the ontological implications of the text. So-called "ontologies", consisting of theories of various kinds expressing the meaning of shared vocabularies, begin to be used for this task. This paper addresses the theoretical issues related to the design and use of such ontologies for purposes of information retrieval and extraction. After a discussion on the nature of semantic matching within a model-theoretical framework, we introduce the subject of Formal Ontology, showing how the notions of parthood, integrity, identity, and dependence can be of help in understanding, organizing and formalizing fundamental ontological distinctions. We present then some basic principles for ontology design, and we illustrate a preliminary proposal for a top-level ontology develped according to such principles. As a concrete example of ontology-based information retrieval, we finally report an ongoing experience of use of a large linguistic ontology for the retrieval of object-oriented software components.

In their paper on "Using Explicit Ontologies in KBS Development", van Heijst and colleagues seem to take for granted Bylander and Chandrasekaran 's hypothesis on the strong dependence of knowledge represesentation on the nature and the inference strategy of the problem at hand, the socalled interaction problem: Representing knowledge for the purpose of solving some problem is strongly affected by the nature of the problem and the inference strategy to be applied to the problem. [Bylander and Chandrasekaran 1988] The fact that the van Heijst and colleagues don't attempt to explore in detail the arguments sustaining this hypothesis is particularly puzzling, since they admit that it contradicts one of the main assumptions of their well-known KADS approach [Schreiber et al. 1993], namely the separation between domain knowledge and problem-solving knowledge. They report two reasons brought by Bylande