This keynote reflects on the value and potential of social networks, interlinked data, semantic web and data-mining. At the same time I would like to elicit new research directions, which are only enabled by the sheer mass of data, sensors, facts, reports, opinions and inter-linkage of people.
knowCube®, a novel multi criteria decision making tool, is introduced. Its user-friendly interface assists “intuitive surfing through decision spaces” by means which are also familiar to non-experts. Causes and effects of alternatives may be examined from different points of view, and several types of criteria – like quantitative or qualitative, dependent or independent, hard or soft, and all mixed together – can be handled at the same time. The tool’s broad applicability is illustrated by some application examples from absolutely different fields: Mixed Sandwiches of various materials are investigated in manufacturing, ideal Foams are produced due to optimal parameter settings, and personalized Drugs could be designed by balancing conflicting effects.
Data Mining describes a technology that discovers non-trivial hidden patterns in a large collection of data. Although this technology has a tremendous impact on our lives, the invaluable contributions of this invisible technology often go unnoticed. This paper addresses the various forms of data mining while providing insights into its expanding role in enriching our life. Emerging forms of data mining are able to perform multidimensional mining on a wide variety of heterogeneous data sources, providing solutions to many problems. This paper highlights the advantages and disadvantages arising from the ever-expanding scope of data mining. Data Mining augments human intelligence by equipping us with a wealth of knowledge and by empowering us to perform our daily task better. As the mining scope and capacity increases, users and organisations become more willing to compromise privacy. The huge data stores of the ‘master miners’ allow them to gain deep insights into individual lifestyles and their social and behavioural patterns. The data on business and financial trends together with the ability to deterministically track market changes will allow an unprecedented manipulation of the stock market. Is it then possible to constrain the scope of mining while delivering the promise of better life?
The evolution of our society to the knowledge based society has raised new challenges for most of the scientific domains that exist. The higher importance given to knowledge extraction instead of getting just information (i.e. data included in a context) hast led to the development of several intelligent techniques for knowledge discovery. This paper shows some examples of using the techniques of case-based reasoning and data-mining for knowledge discovery in the knowledge management system of an university. We have taken as example, the educational domain with the particular case of universities as they represent good examples of organizations that acquire, generate, store and use knowledge for various purposes, teaching, learning and research.
Digital media are used to facilitate social structures thus building digital social networks. Disturbances in such networks occur on different levels (egocentric level, subgroup level, network) and have to be analyzed in the multidimensional context of reference disciplines like sociology and knowledge management. This paper presents a first repository of disturbance patterns for the analysis of digital social networks. Based on the Actor-Network Theory and the Social Network Analysis, new socio-theoretical models for handling complex media settings were developed. On these models a pattern language is defined to describe multidimensional disturbance patterns and to store them in a newly developed pattern repository. The core of the pattern language is the formal expression language for pattern (FELP) which used to specify the structural and the content-specific properties of digital social networks. Results can be visualized with open source graph visualization software. To evaluate the approach a case study has been performed in a repository containing 118 mailing lists and 17.359 individuals. Patterns like troll, spammer and burst have been applied successfully.
This paper contributes to a facet from the area of Web Information Retrieval that has recently received much attention: The satisfaction of a user’s personal information need with respect to text type, presentation type, or information quality. We imply that such properties can be quantified for all kinds ofWeb documents, and we subsume them under the term “Web genre” or “genre”.
Recent surveys show that there is, to a certain degree, a common understanding of Web genre. However, the strictness by which genre and non-genre aspects of a document are experienced is an individual matter. To get a better understanding of the challenges of Web genre identification and its possible limits we investigate in this paper a very interesting question, which has not been posed by now: Given a categorization C of documents (or bookmarks, links, document identifiers), can we provide a reliable assessment whether C is governed by topic or by genre considerations? We present instruments to answer this question as well as to make a distinct statement about the homogeneity of a categorization.
This paper presents some experiments in clustering homogeneous XML documents to validate an existing classification or more generally an organisational structure. Our approach integrates techniques for extracting knowledge from documents with unsupervised classification (clustering) of documents. We focus on the feature selection used for representing documents and its impact on the emerging classification. We mix the selection of structured features with fine textual selection based on syntactic characteristics. We illustrate and evaluate this approach with a collection of Inria activity reports for the year 2003. The objective is to cluster projects into larger groups (Themes), based on the keywords or different chapters of these activity reports. We then compare the results of clustering using different feature selections, with the official theme structure used by Inria.
The discovery of relevant knowledge resources is a remaining problem in large enterprises, where the same problems are often addressed in different locations. In this paper, we propose an enabling infrastructure, which can succcessfully help in discovering personally relevant learning resources (e. g. electronic documents, colleagues, seminars). By using our mobile gotchi framework (MGF), matching between learning requirements and learning resources is improved. The basic idea is autonomous information exchange between agents acting on behalf of their users to proactively find appropriate resources to support daily problem-solving and learning. By sharing personal profiles in an enterprise-specific ontology network, which is autonomously updated, corporate knowledge flows are more transparently represented.
Recent trends in structure and content of global knowledge spaces present new challenges to the field of Knowledge Discovery. Very large, highly structured repositories are increasingly replacing smaller, flat information spaces. Such repositories are often filled with multimedia documents, including image, audio and video data. This publication briefly outlines the underlying trends and discusses implications on approaches to Knowledge Discovery. Some examples for applications accomodating these implications are presented and analysed for lessons learned which can be incorporated in designing future Knowledge Discovery systems. Emphasis is given to the visualisation of hierarchical structures and to cross-media knowledge mining, two fields crucial for adressing future challenges to Knowledge Discovery.
This paper presents an approach to supporting the exchange of knowledge in communities of practice that connect experts from different fields of expertise. The developed system allows unobtrusive construction of personalised knowledge maps that capture implicit knowledge of individuals and groups of users and make it usable for collaborative structuring of shared information repositories. The personalised maps also reflect the global patterns of knowledge exchange in the community which allows the extraction of a shared conceptual structure that connects knowledge across different individuals and groups of users. To this end techniques for self-organised clustering are combined with methods for supervised learning and collaborative filtering. Application scenarios include automatic generation of personalised knowledge portals, collaborative knowledge management and the construction of shared ontologies and topic networks. The real-world testbed is the Internet platfom netzspannung.org.