T1: Knowledge Representation and Reasoning for the Semantic Web - OWL 2 and Rules

Pascal Hitzler, Wright State University, Dayton, Ohio; Markus Krötzsch, Sebastian Rudolph, University of Karlsruhe (TH)

The revision 2 of the Web Ontology Language OWL is much richer than its predecessor OWL 1.0 with respect to modelling with rules. In particular, a significant portion of OWL 2 DL is already expressible using rules (called SROIQ Rules). The tractable profile OWL 2 EL can be extended by rules - within OWL 2 DL - while retaining tractability. Further rules lying outside OWL 2 DL, in particular a generalization of DL-safe Datalog rules, can further be added while still retaining tractability, resulting in a language called ELP, which covers all three tractable profiles of OWL 2. ELP in turn can be extended by local closed-world reasoning such that data complexity still remains polynomial.

This tutorial introduces OWL 2 and the abovementioned rules fragments and extensions in detail. It is aimed for the theoretician as well as the ontology engineer who would like to learn about the intimate relationship between OWL 2 and rules.

Tutorial homepage: http://semantic-web-book.org/page/KI_2009_Tutorial


The presenters are the acknowledged authors of the first German-language textbook, with Springer, on the Semantic Web. They furthermore have an international standing as researchers on the topic. They are actively involved in the W3C standardization working groups for the OWL 2 revision of the Web Ontology Language and for defining the Rule Interchange Format RIF. For more information, please see http://logic.aifb.uni-karlsruhe.de/.


T2: Building Intelligent Mashups

Adrian Giurca, TU Cottbus
Emilian Pascalau, HPI, Potsdam

What do you know about mashups or hybrid web application? Have you tried to create a Web mashup ever? Are you comfortable with editors such as Google Mashup Editor, Microsoft Popfly, Mozilla Ubiquity or Yahoo pipes? Do they still make you feel a bit like a foreign language? Are you looking for more user-interaction in mashups creation and maintenance? Are you looking for more "elegance" for your mashups?

In this tutorial, we will cover many essential intelligent mashup techniques, adding immediately useful tools to your belt. Rationale will be provided for all idioms - the "why" in addition to the "what & how". We will run through lots of small, practical, hands-on examples and will take questions throughout the tutorial.


Adrian Giurca (http://www.informatik.tu-cottbus.de/~agiurca/ ) is currently investigating methods and applications for information systems of the next generation, especially (agent based) reasoning on social media/software and reasoning on the Web and Semantic Web. Formerly, he worked on knowledge representation (particularly with uncertainty representation) and logic programming. Since 2005, he is a senior researcher at the Institute for Informatics of the Brandenburg University of Technology, Germany.


Emilian Pascalau (http://bpt.hpi.uni-potsdam.de/Public/EmilianPascalau ) is a Ph.D. student researching development of methods and enterprise systems comprising browser based rule engines, rules in the context of mashups, and business processes on the Web.

T3: Data Mining for Web 2.0

Andreas Hotho, University of Kassel
Gerd Stumme, University of Kassel

The Web 2.0 is a major source of new data investigated by different research communities. This tutorial will focus on the folksonomies and social bookmarking part of the web 2.0. It begins with a formalization of folksonomies as hypergraphs, which results in two major challenges: 1/ the problem of coping with the enormous size of the available data, due to the large number of users of systems such as del.icio.us, 2/ the complex and rich structure of hypergraphs, requiring the introduction of new graph measures as well as new strategies for data analysis.

To learn more about the structure of folksonomies, we present methods to analyze hypergraphs based on standard graph measures, suitably adapted to hypergraphs, as well as projections of the folksonomy onto simpler graph structures. We focus on clustering measures and on the analysis of tag co-occurrence graphs. In a next step, we show how it is possible to introduce several notions of similarity between the nodes of a folksonomy (resources, users, tags) and how such measures can be used to mine for structures in the folksonomy. In particular, we show how clusters of resources can be identified and characterized. Finally, we present solutions for practical problems like the ranking of resources and tag recommendations in folksonomies.


Andreas Hotho holds a Ph.D. from the University of Karlsruhe, where he worked from 1999 to 2004 at the Institute AIFB. He earned his Master's Degree in information systems at the University of Braunschweig (Germany) in 1998. Since 2004, he is a senior researcher at the University of Kassel. His focus is on the combination of machine learning/data mining and semantic web and also on the analysis of Web 2.0 data.


Gerd Stumme is heading the Hertie Chair on Knowledge & Data Engineering at the University of Kassel and is a member of the Research Center L3S since 2004. He earned his Ph.D. in 1997 at Darmstadt University of Technology, and his habilitation at the Institute AIFB of the University of Karlsruhe in 2002. In 1999/2000, he was Visiting Professor at the University of Clermont-Ferrand, France, and Substitute Professor for Machine Learning and Knowledge Discovery at the University of Magdeburg in 2003.

T4: Humanoid Robots

Sven Behnke, University of Bonn

Humanoid robots, robots with an anthropomorphic body plan and human-like senses, are enjoying increasing popularity as research tools. More and more groups worldwide work on issues such as bipedal locomotion, dexterous manipulation, audio-visual perception, human-robot interaction, adaptive control, and learning, targeted for the application in humanoid robots. These efforts are motivated by the vision to create a new kind of tool: robots that work in close cooperation with humans in the same environment that we designed to suit our needs. While highly specialized industrial robots are successfully employed in industrial mass production, these new applications require a different approach: general purpose humanoid robots.  The human body is well suited for acting in our everyday environments. Stairs, door handles, tools, and so on are designed to be used by humans. A robot with a human-like body can take advantage of these human-centered designs. The new applications will require social interaction between humans and robots. If a robot is able to analyze and synthesize speech, eye movements, mimics, gestures, and body language, it will be capable of intuitive communication with humans. Most of these modalities require a human-like body plan. A human-like action repertoire also facilitates the programming of the robots by demonstration and the learning of new skills by imitation of humans, because there is a one-to-one mapping of human actions to robot actions. Last, but not least, humanoid robots are used as a tool to understand human intelligence. In the same way biomimetic robots have been built to understand certain aspects of animal intelligence, humanoid robots can be used to test models of aspects of human intelligence.

The tutorial is intended to provide support to scientists and practitioners wishing to enter this fascinating area. It introduces the general concepts and principles that underlie humanoid robot systems. A selection of examples provide in-depth practical knowledge on the design of humanoid robot systems. Covered aspects include mechatronic design, perception, behavior control, and learning. Application areas include humanoid robot soccer, intuitive multimodal communication, and mobile manipulation.


Sven Behnke received his M.Sc. degree in Computer Science (Dipl.-Inform.) in 1997 at Martin-Luther-Universität Halle-Wittenberg. In 2002, he obtained a Ph.D. in Computer Science (Dr. rer. nat.) at Freie Universität Berlin. He spent the year 2003 as postdoctoral researcher at the International Computer Science Institute, Berkeley, CA. From 2004 to 2008, Professor Behnke headed the Humanoid Robots Group at Albert-Ludwigs-Universität Freiburg. Since April 2008, he is a  full professor for Autonomous Intelligent Systems at the University of Bonn. His research interests include biologically inspired information processing, humanoid robots, computer vision, speech processing, and machine learning.

T5: Hybrid Planning - Theory and Applications

Susanne Biundo, Bernd Schattenberg, Ulm University

Hybrid Planning is a technique most relevant for real-world applications in areas such as disaster relief, mission control, work-flow management and logistics. It combines decomposition-based planning, where complex tasks are stepwise refined into plans containing simpler subtasks until sequences of executable actions are reached, with explicit reasoning about causality, time and resources. This way, predefined solution-patterns can not only be instantiated and used to solve a current planning problem, they also can be modified substantially and safely be adapted to a current situation. Furthermore, even completely new solutions can be generated from scratch.

The tutorial introduces the methodology of hybrid planning. It presents the formal background of a well-founded integration of task and state hierarchies together with a framework to define a variety of (combinations of) plan generation methods and strategies. Examples from various real-world applications will show how the hybrid planning paradigm enables the adequate modelling of planning domains. Practical planning is demonstrated using the PANDA hybrid planning environment.


Susanne Biundo received her Ph.D. (Dr. rer. nat.) from the University of Karlsruhe. She held a Senior Research Position at the Intelligent User Interfaces Lab of the German Research Center for Artificial Intelligence (DFKI). Since 1998, she is a Professor of Computer Science at Ulm University. Among other professional activities, she served as the Conference and Program Co-chair of various international conferences including KI 2004 and ICAPS 2005. From 1998 to 2003, she headed PLANET, the "European Network of Excellence in AI Planning". Since 2009, she is the chair of the Transregional Collaborative Research Centre on Companion-Technology for Cognitive Technical Systems. Her research interests are in AI Planning, Automated Reasoning, Knowledge Modelling and Cognitive Systems.


Bernd Schattenberg received his M.Sc. degree in Computer Science (Dipl.-Inform.) at Ulm University in 1998. He is a research assistant at the Institute of Artificial Intelligence at the Faculty of Engineering and Computer Science at Ulm University. His main research interests are in Intelligent Planning and Scheduling and their application to complex real-world scenarios.