Lecture Notes in Artificial Intelligence 10505
Subseries of Lecture Notes in Computer Science
LNAI Series Editors Randy Goebel University of Alberta, Edmonton, Canada Yuzuru Tanaka Hokkaido University, Sapporo, Japan Wolfgang Wahlster DFKI and Saarland University, Saarbrücken, Germany
LNAI Founding Series Editor Joerg Siekmann DFKI and Saarland University, Saarbrücken, Germany More information about this series at http://www.springer.com/series/1244 Gabriele Kern-Isberner • Johannes Fürnkranz Matthias Thimm (Eds.)
KI 2017: Advances in Artificial Intelligence 40th Annual German Conference on AI Dortmund, Germany, September 25–29, 2017 Proceedings
123 Editors Gabriele Kern-Isberner Matthias Thimm Fakultätfür Informatik FB Informatik Technische Universität Dortmund Universität Koblenz Dortmund Koblenz, Rheinland-Pfalz Germany Germany Johannes Fürnkranz FB Informatik TU Darmstadt Darmstadt, Hessen Germany
ISSN 0302-9743 ISSN 1611-3349 (electronic) Lecture Notes in Artificial Intelligence ISBN 978-3-319-67189-5 ISBN 978-3-319-67190-1 (eBook) DOI 10.1007/978-3-319-67190-1
Library of Congress Control Number: 2017953421
LNCS Sublibrary: SL7 – Artificial Intelligence
© Springer International Publishing AG 2017 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Printed on acid-free paper
This Springer imprint is published by Springer Nature The registered company is Springer International Publishing AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland Preface
The German conference on Artificial Intelligence (abbreviated KI for “Künstliche Intelligenz”) looks back on a long and fruitful history. The first official event took place in 1975, at that time a workshop of the KI working group of the German “Gesellschaft für Informatik” (association for computer science, GI). Before that, there were inofficial meetings, such as the “Fachtagung Kognitive Verfahren und Systeme”, which was held in Hamburg in April 1973. The meeting has now developed into an annual conference for researchers in artificial intelligence, primarily from Germany and its neighboring countries but open to international participation. This volume contains the papers presented at the 40th event in this series, which was held at the Technical University of Dortmund, September 25–29th, 2017. This year we received 73 valid submissions, an increase of 50% over last year. We were able to accept 20 papers as full research papers and 16 as short technical communications, yielding an acceptance rate of 27% for full papers and 49% overall. Due to the limited number of available slots in the conference schedule, we had to make difficult decisions and several worthy submissions had to be rejected or downgraded from full to short papers. The Program Committee worked very hard to thoroughly review all the submitted papers and to provide action points to improve the papers. Despite the increased workload for the PC, almost all papers received three reviews, and only 10 papers had to be selected or rejected on the basis of only 2 reviews. The program chairs managed discussions amongst the reviewers, from which the final decisions emerged. As a result, the contributions cover a range of topics from, e.g., agents, robotics, cognitive sciences, machine learning, planning, knowledge representation, reasoning, and ontologies, with numerous applications in areas like social media, psychology, and transportation sys- tems, reflecting the richness and diversity of our field. In addition to the regular sessions, our program also featured three invited talks by Pierre Baldi (University of California, Irvine), Gerhard Brewka (University of Leipzig), and Luc De Raedt (Katholieke Universiteit Leuven), as well as an industrial session featuring a keynote by Wolfgang Wahlster (Saarland University), and a session com- posed of presentations of selected papers by German authors that have been presented at our international sister conferences AAAI and IJCAI in 2017. In order to celebrate the 40th anniversary of the German Conference on Artificial Intelligence, the program also contained a historical session with a panel discussion. The session was hosted by Ulrich Furbach, and contained contributions from the panelists Katharina Morik, Hans-Helmut Nagel, Bernd Neumann, and Jörg Siekmann. After a short review of the history of computer science and artificial intelligence in Germany by the host, the panelists commented on their early AI-related activities and the relationship of these activities with the computer science community at that time. Finally, the development of the field in recent years and its future was reflected in an open forum. VI Preface
For the first two days of the conference, our workshop and tutorial chair, Christoph Beierle (University of Hagen), organized a program of two workshops: – ZooOperation Competition (Vanessa Volz, Christian Eichhorn) – Formal and Cognitive Reasoning (Christoph Beierle, Gabriele Kern-Isberner, Marco Ragni, Frieder Stolzenburg) During the workshops, many additional papers were presented, ideas discussed, and experiences exchanged. Moreover, the program also featured two tutorials: – Defeasible Reasoning for Description Logics (Ivan Varzinczak) – Knowledge Representation and Reasoning with Nilsson-Style Probabilistic Logics (Nico Potyka) Organizing such a traditional conference is a very challenging but no less rewarding experience, which would not have been possible without the help of the many indi- viduals who contributed to the success of this event. First and foremost, we would like to thank the authors and the reviewers for their excellent work, which forms the core of any such meeting. We also thank our workshop and tutorial chair, the invited speakers, the workshop chairs and tutorial presenters, and the participants of the historical ses- sion, all of which have already been listed above. Last but not least, special thanks go to the local organization team from the Technical University of Dortmund, Christian Eichhorn, Steffen Schieweck, and Marco Wilhelm without whom this con- ference would not have been possible.
July 2017 Gabriele Kern-Isberner Johannes Fürnkranz Matthias Thimm Organization
General Chair
Gabriele Kern-Isberner TU Dortmund, Germany
Program Chairs
Johannes Fürnkranz TU Darmstadt, Germany Matthias Thimm University of Koblenz-Landau, Germany
Workshop and Tutorial Chair
Christoph Beierle University of Hagen, Germany
Historial Session Chair
Ulrich Furbach University of Koblenz-Landau, Germany
Local Chairs
Christian Eichhorn TU Dortmund, Germany Steffen Schieweck TU Dortmund, Germany Marco Wilhelm TU Dortmund, Germany
Program Committee
Klaus-Dieter Althoff DFKI/University of Hildesheim, Germany Franz Baader TU Dresden, Germany Christian Bauckhage Fraunhofer IAIS, Germany Sven Behnke University of Bonn, Germany Christoph Beierle University of Hagen, Germany Ralph Bergmann University of Trier, Germany Leopoldo Bertossi Carleton University, Germany Chris Biemann University of Hamburg, Germany Gerhard Brewka Leipzig University, Germany Philipp Cimiano Bielefeld University, Germany Jürgen Dix Clausthal University of Technology, Germany Igor Douven Paris-Sorbonne University, France Didier Dubois IRIT/RPDMP, France Stefan Edelkamp University of Bremen, Germany Christian Guttmann Nordic AI Institute/Karolinska Institute/ Univ. of New South Wales, Australia VIII Organization
Barbara Hammer Bielefeld University, Germany Malte Helmert University of Basel, Switzerland Andreas Hotho University of Würzburg, Germany Eyke Hüllermeier University of Paderborn, Germany Anthony Hunter University College London, UK Steffen Hölldobler TU Dresden, Germany Dietmar Jannach TU Dortmund, Germany Jean Christoph Jung Universität Bremen, Germany Kristian Kersting TU Darmstadt, Germany Oliver Kramer Universität Oldenburg, Germany Ralf Krestel Hasso Plattner Institute/University of Potsdam, Germany Thomas Lukasiewicz University of Oxford, UK Till Mossakowski University of Magdeburg, Germany Maurice Pagnucco The University of New South Wales, Australia Heiko Paulheim University of Mannheim, Germany Rafael Peñaloza Free University of Bozen-Bolzano, Italy Giuseppe Pirrò Institute for High Performance Computing and Networking (ICAR-CNR), Italy Henri Prade IRIT/CNRS, France Stefan Roth TU Darmstadt, Germany Günter Rudolph TU Dortmund, Germany Sebastian Rudolph TU Dresden, Germany Klaus-Dieter Schewe Software Competence Center Hagenberg, Germany Ute Schmid University of Bamberg, Germany Lars Schmidt-Thieme University of Hildesheim, Germany Lutz Schröder Friedrich-Alexander-Universität Erlangen-Nürnberg, Germany Christoph Schwering University of New South Wales, Australia Steffen Staab University Koblenz-Landau/Univ. of Southampton, UK Hannes Strass Leipzig University, Germany Heiner Stuckenschmidt University of Mannheim, Germany Thomas Stützle Université Libre de Bruxelles, Germany Paul Thorn Heinrich-Heine-UniversitätDüsseldorf, Germany Ingo J. Timm University of Trier, Germany Anni-Yasmin Turhan TU Dresden, Germany Jilles Vreeken MPI for Informatics/Saarland University, Germany Toby Walsh NICTA/University of New South Wales, Australia Stefan Woltran TU Wien, Vienna Organization IX
Additional Reviewers
Ahlbrecht, Tobias Koo, Seongyong Apeldoorn, Daan Kumar, Abhishek Asaadi, Shima Kutsch, Steven Ayzenshtadt, Viktor Ludewig, Malte Barteld, Fabian Ma, Ning Bliem, Bernhard Martin Garcia, German Borgwardt, Stefan Martinez-Gil, Jorge Bremer, Jörg Milde, Benjamin Buga, Andreea Möller, Ralf Dang, Hien Müller, Gilbert Dietz, Emmanuelle-Anna Nemes, Tania Ecke, Andreas Nesi, Monica Farazi, Hafez Neuhaus, Fabian Fiekas, Niklas Nunes, Ingrid Flesca, Sergio Plaza, Enric Grumbach, Lisa Pommerening, Florian Hassan, Teena Chakkalayil Sauerwald, Kai Henzgen, Sascha Schweizer, Lukas Houben, Sebastian Seuß, Dominik Huber, Steffen Siebers, Michael Igarashi, Ayumi Stram, Rotem Jugovac, Michael Wernhard, Christoph Keller, Thomas Wirth, Christian Knees, Peter Zeyen, Christian Keynotes Deep Learning: Theory, Algorithms, and Applications in the Natural Sciences
Pierre Baldi
University of California, Irvine (UCI) [email protected]
Abstract. The process of learning is essential for building natural or artificial intelligent systems. Thus, not surprisingly, machine learning is at the center of artificial intelligence today. And deep learning—essentially learning in complex systems comprised of multiple processing stages—is at the forefront of machine learning. In the last few years, deep learning has led to major performance advances in a variety of engineering disciplines from computer vision, to speech recognition, to natural language processing, and to robotics. In this talk we will first address some fundamental theoretical issues about deep learning through the theory of local learning and deep learning channels. We will then describe inner and outer algorithms for designing deep recursive neural architectures to process structured, variable-size, data such as biological or natural language sequences, phylogenetic or parse trees, and small or large molecules in biochemistry. Finally we will present various applications of deep learning to problems in the natural sciences, such as the detection of exotic particles in high-energy physics, the prediction of molecular properties and reactions in chemistry, and the prediction of protein structures in biology. Computational Models of Argument: A Fresh View on Old AI Problems
Gerhard Brewka
Department of Computer Science, Leipzig University, Germany [email protected]
Abstract. In the last two decades symbolic AI has seen a steady rise of interest in the notion of argument, an old topic of study in philosophy. This interest was fueled by a certain dissatisfaction with existing approaches in knowledge rep- resentation, especially default reasoning and inconsistency handling, and by the demands of applications in legal reasoning and related fields. The ultimate goal of computational argumentation is to enable the development of computer-based systems capable to support and to participate in argumentative activities. To this end one has to come up with formal models of the way we usually come to conclusions and make decisions, namely by 1. constructing arguments for and against various options, 2. establishing relationships among the arguments, most notably the attack relation, and 3. identifying interesting subsets of the arguments which represent coherent positions based on these relations. In the talk we will highlight some of the main ideas and key techniques that have been developed in the field and show how they provide new ways of representing knowledge, handling inconsistencies, and reasoning by default. In particular, we will demonstrate how directed graphs with arbitrary edge labels, which are widely used to visualize argumentation and reasoning scenarios, can be turned into full-fledged knowledge representation formalisms with a whole range of precisely defined semantics. Probabilistic Programming and its Applications
Luc De Raedt
Department of Computer Science, Katholieke Universiteit Leuven, Belgium [email protected]
Abstract. Probabilistic programs combine the power of programming languages with that of probabilistic graphical models. There has been a lot of progress in this paradigm over the past twenty years. This talk will introduce probabilistic logic programming languages [1], which are based on Sato's distribution semantics and which extend probabilistic databases. The key idea is that facts or tuples can be annotated with probabilities that indicate their degree of belief. Together with the rules that encode domain knowledge they induce a set of possible worlds. After an introduction to probabilistic programs, which will cover semantics, inference, and learning, the talk will sketch some emerging applications in knowledge based systems, in cognitive robotics and in answering probability questions. The first is concerned with learning rules in knowledge based systems such as CMU's Never Ending Language Learning [2], the second with learning probabilistic action definitions and using these for planning to grasp certain objects [3], the final one with the answering of challenging mathematical exercises about probability that are formulated in natural language [4].
References
1. De Raedt, L., Kimmig, A.: Probabilistic (logic) programming concepts. Mach. Learn. 100(1), 5–47 (2015) 2. De Raedt, L., Dries, A., Thon, I., Van den Broeck, G., Verbeke, M.: Inducing probabilistic relational rules from probabilistic examples. In: Proceedings of the 25th International Joint Conference on Artificial Intelligence, IJCAI-15, pp. 1835–1843 (2015) 3. Nitti, D., Ravkic, I., Davis, J., De Raedt, L.: Learning the structure of dynamic hybrid relational models. In: Proceedings of the 22nd European Conference on Artificial Intelligence, ECAI-16, pp. 1283–1290 (2016) 4. Dries, A., Kimmig, A., Davis, J., Belle, V., De Raedt, L.: Solving probability problems in natural language. In: Proceedings of the 26th International Joint Conference on Artficial Intelligence, IJCAI-17, pp. 3981–3987 (2017) Artificial Intelligence for Industrie 4.0
Wolfgang Wahlster
DFKI and Saarland University www.dfki.de/*wahlster
Abstract. The transformative power of Artificial Intelligence (AI) for the fourth industrial revolution based on cyber-physical production systems is now rec- ognized globally by highly industrialized nations. When we coined the term Industrie 4.0 in 2010, it was already clear to me that machine learning, semantic technologies, real-time action planning as well as plan recognition, collaborative robotics, and intelligent user interfaces are the scientific foundation for smart factories, smart products and smart services. AI is a key enabler for the next generation of smart manufacturing in Industrie 4.0, since it leads to a disruption in traditional workflows, supply chains, value creation, and business models in manufacturing and works towards empowering and expanding workforce expertise. The use of AI in manufacturing is paving the way to the synergistic collaboration between humans and robots in urban smart factories for mass customization [2]. In particular, we present recent results from our Industrie 4.0 projects at DFKI, including hybrid teams of human workers and collaborative robots, deep learning for predictive maintenance of networked production machines and for understanding human behaviors of shop floor workers, semantic technologies for worldwide interoperability of machine-to-machine communication in smart factories and logistics, human-aware and real-time production planning and scheduling for multiagent systems, intelligent industrial assistance systems for human workers, and proactive and situation-aware on-line help and training on the shop floor. The concept of active semantic product memories [3] that serve as digital twins invert the traditional production logic, since in Industrie 4.0 the emerging product is controlling its own pro- duction process in a service-oriented multiagent architecture. We discuss use cases from legacy factories which we have upgraded to Industrie 4.0 and show the comparative gains [1] in productivity, stock reduction, resource efficiency, retooling or changeover times, and job satisfaction.
References
1. Schuh, G., Anderl, R., Gausemeier J., ten Hompel, M., Wahlster, W. (eds.) Industrie 4.0 Maturity Index. Managing the Digital Transformation of Companies. Munich: Herbert Utz. 2. Wahlster, W.: Semantic technologies for mass customization. In: Wahlster, W., Grallert, HJ., Wess, S., Friedrich, H., Widenka, T. (eds.) Towards the Internet of Services, pp. 3–14. Springer, Heidelberg (2014) 3. Wahlster, W.: The semantic product memory: an interactive black box for smart objects. In: Wahlster (ed.) SemProM: Foundations of Semantic Product Memories for the Internet of Things, pp. 3–21. Springer, Heidelberg (2013) Contents
Full Technical Papers
Employing a Restricted Set of Qualitative Relations in Recognizing Plain Sketches ...... 3 Ahmed M.H. Abdelfattah and Wael Zakaria
Interval Based Relaxation Heuristics for Numeric Planning with Action Costs ...... 15 Johannes Aldinger and Bernhard Nebel
Expected Outcomes and Manipulations in Online Fair Division ...... 29 Martin Aleksandrov and Toby Walsh
Most Competitive Mechanisms in Online Fair Division ...... 44 Martin Aleksandrov and Toby Walsh
A Thorough Formalization of Conceptual Spaces ...... 58 Lucas Bechberger and Kai-Uwe Kühnberger
Propagating Maximum Capacities for Recommendation ...... 72 Ahcène Boubekki, Ulf Brefeld, Cláudio Leonardo Lucchesi, and Wolfgang Stille
Preventing Groundings and Handling Evidence in the Lifted Junction Tree Algorithm ...... 85 Tanya Braun and Ralf Möller
Improving the Cache-Efficiency of Shortest Path Search ...... 99 Stefan Edelkamp
Automating Emendations of the Ontological Argument in Intensional Higher-Order Modal Logic...... 114 David Fuenmayor and Christoph Benzmüller
Real-Time Public Transport Delay Prediction for Situation-Aware Routing. . . 128 Lukas Heppe and Thomas Liebig
Online Multi-object Tracking-by-Clustering for Intelligent Transportation System with Neuromorphic Vision Sensor ...... 142 Gereon Hinz, Guang Chen, Muhammad Aafaque, Florian Röhrbein, Jörg Conradt, Zhenshan Bing, Zhongnan Qu, Walter Stechele, and Alois Knoll XVIII Contents
A Generalization of Probabilistic Argumentation with Dempster-Shafer Theory ...... 155 Nguyen Duy Hung
Evolving Kernel PCA Pipelines with Evolution Strategies ...... 170 Oliver Kramer
An Experimental Study of Dimensionality Reduction Methods ...... 178 Almuth Meier and Oliver Kramer
Planning with Independent Task Networks ...... 193 Felix Mohr, Theo Lettmann, and Eyke Hüllermeier
Complexity-Aware Generation of Workflows by Process-Oriented Case-Based Reasoning...... 207 Gilbert Müller and Ralph Bergmann
LiMa: Sequential Lifted Marginal Filtering on Multiset State Descriptions . . . 222 Max Schröder, Stefan Lüdtke, Sebastian Bader, Frank Krüger, and Thomas Kirste
A Priori Advantages of Meta-Induction and the No Free Lunch Theorem: A Contradiction?...... 236 Gerhard Schurz and Paul Thorn
Dynamic Map Update of Non-static Facility Logistics Environment with a Multi-robot System ...... 249 Nayabrasul Shaik, Thomas Liebig, Christopher Kirsch, and Heinrich Müller
Similarity and Contrast on Conceptual Spaces for Pertinent Description Generation ...... 262 Giovanni Sileno, Isabelle Bloch, Jamal Atif, and Jean-Louis Dessalles
Technical Communications
Alarm Management on a Liquid Bulk Terminal ...... 279 Bram Aerts, Kylian Van Dessel, and Joost Vennekens
Action Model Acquisition Using Sequential Pattern Mining ...... 286 Ankuj Arora, Humbert Fiorino, Damien Pellier, and Sylvie Pesty
Steering Plot Through Personality and Affect: An Extended BDI Model of Fictional Characters...... 293 Leonid Berov Contents XIX
Ontological Modelling of a Psychiatric Clinical Practice Guideline ...... 300 Daniel Gorín, Malte Meyn, Alexander Naumann, Miriam Polzer, Ulrich Rabenstein, and Lutz Schröder
A Logic Programming Approach to Collaborative Autonomous Robotics . . . . 309 Binal Javia and Philipp Cimiano
Parametrizing Cartesian Genetic Programming: An Empirical Study ...... 316 Paul Kaufmann and Roman Kalkreuth
Gesture ToolBox: Touchless Human-Machine Interface Using Deep Learning...... 323 Elann Lesnes-Cuisiniez, Jesus Zegarra Flores, and Jean-Pierre Radoux
Analysis of Sound Localization Data Generated by the Extended Mainzer Kindertisch ...... 330 Daniel Lückehe, Katharina Schmidt, Karsten Plotz, and Gabriele von Voigt
A Robust Number Parser Based on Conditional Random Fields ...... 337 Heiko Paulheim
From Natural Language Instructions to Structured Robot Plans ...... 344 Mihai Pomarlan, Sebastian Koralewski, and Michael Beetz
Opportunistic Planning with Recovery for Robot Safety...... 352 Bernhard Reiterer and Michael Hofbaur
Towards Simulation-Based Role Optimization in Organizations ...... 359 Lukas Reuter, Jan Ole Berndt, and Ingo J. Timm
One Knowledge Graph to Rule Them All? Analyzing the Differences Between DBpedia, YAGO, Wikidata & co...... 366 Daniel Ringler and Heiko Paulheim
ClassifyHub: An Algorithm to Classify GitHub Repositories ...... 373 Marcus Soll and Malte Vosgerau
Bremen Big Data Challenge 2017: Predicting University Cafeteria Load . . . . 380 Jochen Weiner, Lorenz Diener, Simon Stelter, Eike Externest, Sebastian Kühl, Christian Herff, Felix Putze, Timo Schulze, Mazen Salous, Hui Liu, Dennis Küster, and Tanja Schultz
Towards Sentiment Analysis on German Literature ...... 387 Albin Zehe, Martin Becker, Fotis Jannidis, and Andreas Hotho
Author Index ...... 395