Contributions published at Dynamic and Distributed Information Systems (Abraham Bernstein)

null

This is the initial technical report of the Semantic Web Services Language (SWSL) Committee of the Semantic Web Services Initiative (SWSI). This report consists of the following four top-level documents, with four related appendices. * Semantic Web Services Framework (SWSF) Overview * The Semantic Web Services Language (SWSL) * The Semantic Web Services Ontology (SWSO) * SWSF Application Scenarios Appendices (of the Ontology document): * PSL in SWSL-FOL and SWSL-Rules * Axiomatization of the FLOWS Process Model * Axiomatization of the Process Model in SWSL-Rules * Reference Grammars

A data mining (DM) process involves multiple stages. A simple, but typical, process might in-clude preprocessing data, applying a data-mining algorithm, and postprocessing the mining re-sults. There are many possible choices for each stage, and only some combinations are valid. Because of the large space and non-trivial interactions, both novices and data-mining specialists need assistance in composing and selecting DM processes. Extending notions developed for statistical expert systems we present a prototype Intelligent Discovery Assistant (IDA), which provides users with (i) systematic enumerations of valid DM processes, in order that important, potentially fruitful options are not overlooked, and (ii) effective rankings of these valid processes by different criteria, to facilitate the choice of DM processes to execute. We use the prototype to show that an IDA can indeed provide useful enumerations and effective rankings in the context of simple classification processes. We discuss how an IDA could be an important tool for knowledge sharing among a team of data miners. Finally, we illustrate the claims with a com-prehensive demonstration of cost-sensitive classification using a more involved process and data from the 1998 KDDCUP competition.

The semantic web presents the vision of a distributed, dynamically growing knowledge base founded on formal logic. Common users, however, seem to have problems even with the simplest Boolean expression. As queries from web search engines show, the great majority of users simply do not use Boolean expressions. So how can we help users to query a web of logic that they do not seem to understand? We address this problem by presenting a natural language front-end to semantic web querying. The front-end allows formulating queries in Attempto Controlled English (ACE), a subset of natural English. Each ACE query is translated into a discourse representation structure – a variant of the language of first-order logic – that is then translated into the semantic web querying language PQL. As examples show, our approach offers great potential for bridging the gap between the semantic web and its real-world users, since it allows users to query the semantic web without having to learn an unfamiliar formal language.

In this paper we introduce the approach of direct interruptability inference from accelerometer and audio data and show that it provides highly accurate and robust predictions. Furthermore, we argue that scenarios are central for evaluating the performance of interruptability predicting devices and prove it on our setup. We also demonstrate that scenarios provide the foundation for avoiding misleading results, assessing the results’ generalizability, and provide the basis for a stratified scenario-based learning model, which greatly speeds-up the training of such devices.

Finding a good similarity assessment algorithm for the use in ontologies is central to the functioning of techniques such as retrieval, matchmaking, clustering, data-mining, ontology translations, automatic database schema matching, and simple object comparisons. This paper assembles a catalogue of ontology based similarity measures, which are experimentally compared with a �similarity gold standard� obtained by surveying 50 human subjects. Results show that human and algorithmic similarity predications varied substantially, but could be grouped into cohesive clusters. Addressing this variance we present a personalized similarity assessment procedure, which uses a machine learning component to predict a subject�s cluster membership, providing an excellent prediction of the gold standard. We conclude by hypothesizing ontology dependent similarity measures.

People are subjected to a multitude of interruptions, which in some situations are detrimental to their work performance. Consequently, the capability to predict a person’s degree of interruptability (i.e., a measure of detrimental an interruption would be to her current work) can provide a basis for a ?ltering mechanism. This paper introduces a novel approach to predict a person’s presence and interruptability in an of?ce-like environment based on audio, multi-sector motion detection using video, and the time of the day collected as sensor data. Conducting an experiment in a real of?ce environment over the length of more than 40 work days we show that the multisector motion detection data, which to our knowledge has been used for the ?rst time to this end, outperforms audio data both in presence and interruptability. We, furthermore, show, that the combination of all three data-streams improves the interruptability prediction accuracy and robustness. Finally, we use these data to predict a subject’s phone behavior (ignore or accept the incoming phone call) by combining interruptability and the estimated importance of call. We call such an application an arti?cial receptionist. Our analysis also show that the results improve when taking the temporal aspect of the context into account.

The semantic web presents the vision of a distributed, dynamically growing knowledge base founded on formal logic. Common users, however, seem to have problems even with the simplest Boolean expression. As queries from web search engines show, the great majority of users simply do not use Boolean expressions. So how can we help users to query a web of logic that they do not seem to understand? We address this problem by presenting a natural language front-end to semantic web querying. The front-end allows formulating queries in Attempto Controlled English (ACE), a subset of natural English. Each ACE query is translated into a discourse representation structure – a variant of the language of first-order logic – that is then translated into the semantic web querying language PQL. As examples show, our approach offers great potential for bridging the gap between the semantic web and its real-world users, since it allows users to query the semantic web without having to learn an unfamiliar formal language.

Finding good algorithms for assessing the similarity of complex objects in ontologies is central to the functioning of techniques such as retrieval, matchmaking, clustering, data-mining, semantic sense disambiguation, ontology translations, and simple object comparisons. These techniques provide the basis for supporting a wide variety of business intelligence computing tasks like finding a process in a best practice repository, finding a suitable service provider/outsourcing partner for agile process enactment, dynamic customer segmentation, semantic web applications, and systems integration. To our knowledge, however, there exists no study that systematically compares the prediction quality of ontology based similarity measures. This paper assembles a catalogue of ontology based similarity measures that are (partially) adapted from related domains. These measures are compared to each other within a large, real-world best practice ontology as well as evaluated in a realistic business process retrieval scenario. We find that different similarity algorithms reflect different notions of similarity. We also show how a combination of similarity measures can be used to improve both precision and recall of an ontology-based, query-by-example style, object retrieval approach. Combining the study’s findings with the literature we argue for the need of extended studies to assemble a more complete catalogue of object similarity measures that can be evaluated in many applications and ontologies.

null

Online repositories are increasingly being called on to provide access to services that describe or provide useful behaviors. Existing techniques for finding the services offer low retrieval precision, returning many irrelevant matches. This article describes a novel service retrieval approach that captures service semantics via process models, and applies a pattern-matching algorithm to locate desired services.

A critical need for many organizations is the ability to quickly (re-)design their business processes in response to changing needs and capabilities. Current process design tools and methodologies, however, are very resource-intensive and provide little support for generating (as opposed to merely recording) new design alternatives. This paper describes the 'process recombination,' a novel approach for template-based business process re-design based upon the MIT Process Handbook. This approach allows one to systematically generate different process (re-) designs using the repository of process alternatives stored in the Process Handbook. Our experience to date has shown that this approach can be effective in helping users produce innovative process designs.

Semantic Web Services (SWS), the convergence of Semantic Web and Web Services, is the emerging next major generation of the Web, in which e-services and business communication become more knowledge-based and agent-based. In the SWS vision, service descriptions are built partly upon process ontologies – widely shared ontological knowledge about business processes – which are represented using Semantic Web techniques for declarative knowledge representation (KR), e.g., OWL Description Logic or RuleML Logic Programs. In this paper, we give the first approach to solving a previously unsolved, crucial problem in representing process ontologies using SW KR: how to represent non-monotonic inheritance reasoning, in which at each (sub)class in the class hierarchy, any inherited property value may be overridden with another value, or simply cancelled (i.e., not inherited). Non-monotonic inheritance is an important, heavily-used feature in pre-SWS process ontologies, e.g., ubiquitous in object-oriented (OO) programming. The advantages of non-monotonicity in inheritance include greater reuse/modularity and easier specification, updating, and merging. We focus in particular on the Process Handbook (PH), a large, influential, and well-used process ontologies repository that is representative in its features for non-monotonic inheritance. W3C’s OWL, the currently dominant SW KR for ontologies, is fundamentally incapable of representing non-monotonicity; so too is First Order Logic. Using instead another form of leading SW KR – RuleML – we give a new approach that successfully represents the PH’s style of non-monotonic inheritance. In this Courteous Inheritance approach, PH ontology knowledge is represented as prioritized default rules expressed in the Courteous Logic Programs (CLP) subset of RuleML. A prototype of our approach is in progress. We aim to use it to enable SWS exploitation of the forthcoming open-source version of the PH.

In the past, most collaboration support systems have focused on either automating fixed work processes or simply supporting communication in ad-hoc processes. This results in systems that are usually inflexible and difficult to change or that provide no specific support to help users decide what to do next. This paper describes a new kind of tool that bridges the gap between these two approaches by flexibly supporting processes at many points along the spectrum: from highly specified to highly unspecified. The development of this approach was strongly based on social science theory about collaborative work.

A critical need for many organizations in the next century will be the ability to quickly develop innovative business processes to take advantage of rapidly changing technologies and markets. Current process design tools and methodologies, however, are very resource-intensive and provide little support for generating (as opposed to merely recording) new design alternatives. This paper describes the Process Recombinator, a novel tool for generating new business process ideas by recombining elements from a richly structured repository of knowledge about business processes. The key contribution of the work is the technical demonstration of how such a repository can be used to automatically generate a wide range of innovative process designs. We have also informally evaluated the Process Recombinator in several field studies, which are briefly described here as well.

This paper describes a novel theoretical and empirical approach to tasks such as business process redesign and knowledge management. The project involves collecting examples of how different organizations perform similar processes, and organizing these examples in an on-line ìprocess handbook"". The handbook is intended to help people: (1) redesign existing organizational processes, (2) invent new organizational processes (especially ones that take advantage of information technology), and (3) share ideas about organizational practices. A key element of the work is an approach to analyzing processes at various levels of abstraction, thus capturing both the details of specific processes as well as the ""deep structure"" of their similarities. This approach uses ideas from computer science about inheritance and from coordination theory about managing dependencies. A primary advantage of the approach is that it allows people to explicitly represent the similarities (and differences) among related processes and to easily find or generate sensible alternatives for how a given process could be performed. In addition to describing this new approach, the work reported here demonstrates the basic technical feasibility of these ideas and gives one example of their use in a field study.

This article investigates how to support process enactment in highly flexible organizations. First it develops the requirements for such a support system. Then it proposes a prototype implementation, which offers its users the equivalent of a CAD/CAM-like tool for designing and supporting business processes. The tool enables end-users to take flexible building blocks of a production process, reassemble them to fit the specific needs of a particular case and finally export its description to process support systems like workflow management systems.

This paper addresses the classification of linked entities. We introduce a relational vector-space (VS) model (in analogy to the VS model used in information retrieval) that abstracts the linked structure, representing entities by vectors of weights. Given labeled data as background knowledge/training data, classification procedures can be defined for this model, including a straightforward, “direct” model using weighted adjacency vectors. Using a large set of tasks from the domain of company affiliation identification, we demonstrate that such classification procedures can be effective. We then examine the method in more detail, showing that as expected the classification performance correlates with the relational autocorrelation of the data set. We then turn the tables and use the relational VS scores as a way to analyze/visualize the relational autocorrelation present in a complex linked structure. The main contribution of the paper is to introduce the relational VS model as a potentially useful addition to the toolkit for relational data mining. It could provide useful constructed features for domains with low to moderate relational autocorrelation; it may be effective by itself for domains with high levels of relational autocorrelation, and it provides a useful abstraction for analyzing the properties of linked data.

This paper addresses the classification of linked entities. We introduce a relational vector-space (VS) model (in analogy to the VS model used in information retrieval) that abstracts the linked structure, representing entities by vectors of weights. Given labeled data as background knowledge/training data, classification procedures can be defined for this model, including a straightforward, “direct” model using weighted adjacency vectors. Using a large set of tasks from the domain of company affiliation identification, we demonstrate that such classification procedures can be effective. We then examine the method in more detail, showing that as expected the classification performance correlates with the relational autocorrelation of the data set. We then turn the tables and use the relational VS scores as a way to analyze/visualize the relational autocorrelation present in a complex linked structure. The main contribution of the paper is to introduce the relational VS model as a potentially useful addition to the toolkit for relational data mining. It could provide useful constructed features for domains with low to moderate relational autocorrelation; it may be effective by itself for domains with high levels of relational autocorrelation, and it provides a useful abstraction for analyzing the properties of linked data.

null