The ECAI 2014 tutorials will be held at the Czech Technical University on August 18-19.

All participants have to register for ECAI conference, it is not possible to attend only tutorials.

Tutorials provide an opportunity for junior and senior researchers to freely explore exciting advances in disciplines outside their normal focus. We believe this type of forum is essential for the cross fertilization, cohesiveness, and vitality of the AI field. We all have a lot to learn from each other. The tutorial programme, consisting of six half-day tutorials on established research areas, and eight 90-minute spotlight tutorials on emerging or more specialized topics, promotes the continuing education of each member of our research community.

To encourage full participation by technical conference registrants, no separate fee will be charged for admittance to the tutorial programme. However, for organization convenience, we kindly ask you to register for all ECAI tutorials you are interested in, in advance.

ECAI participants will also have access to a number of other tutorials at ECAI workshops and co-located events.

List of Tutorials

Time schedule HERE.

Monday, August 18, 9:00–12:30

  •  T1 : Constraint Processing: From Algorithms to Applications
    Roman Barták
  • T2: Temporal Representation and Reasoning in Interval Temporal Logics
    Angelo Montanari, Pietro Sala and Dario Della Monica

Monday, August 18, 9:00–10:30

  • T3: The Yesterday, Today, and Tomorrow of Parallelism in Declarative Programming
    Enrico Pontelli

Monday, August 18, 11:00–12:30

  • T4: Metalearning & Algorithm Selection
    Pavel Brazdil, Carlos Soares and Joaquin Vanschoren

Monday, August 18, 14:00–17:30

  • T5: Software Engineering for Search and Optimization Problems
    Luca Di Gaspero and Tommaso Urli
  • T6: Formal Methods for Event Processing
    Alexander Artikis and George Paliouras

Monday, August 18, 14:00–15:30

  • T7: Tabling for Planning
    Neng-Fa Zhou

Monday, August 18, 16:00–17:30

  • T8: Probabilistic Programming
    Luc De Raedt and Angelika Kimmig

Tuesday, August 19, 9:00–12:30

  • T9: Introduction to Statistical and Behavioural Analysis of Agent-Based Models
    Tamás Máhr, László Gulyás and Richárd O. Legéndi

Tuesday, August 19, 9:00–10:30

  • T10: Procedural Content Generation in Games
    Noor Shaker

Tuesday, August 19, 11:00–12:30

  • T11: Replication and Recomputation in Scientific Experiments
    Ian Gent and Lars Kotthoff

Tuesday, August 19, 14:00–17:30

  • T12: Search Methods for Classical and Temporal Planning
    Jussi Rintanen

Tuesday, August 19, 14:00–15:30

  • T13: Intelligent Socio Technical Interaction
    Virginia Dignum and Frank Dignum

Tuesday, August 19, 16:00–17:30

  • T14: Multilingual Semantic Processing with BabelNet
    Roberto Navigli and Andrea Moro

Other interesting tutorials at ECAI

Tuesday, August 19, 9:00-10:30

  • TR1 Rule Learning (at RuleML) 
    Johannes Fuernkranz

AUGUST 18, Morning

T1: Constraint Processing: From Algorithms to Applications
(Half-day, 9:00–12:30)

Roman Barták


Constraint Programming (CP) is a technology for solving combinatorial optimisation problems in areas such as planning, scheduling, and assignment problems, circuit design, network management and configuration, interactive graphics, molecular biology etc. A user specifies the problem in a declarative way by using decision variables and constraints restricting allowed combinations of values to be assigned to the variables. Then, the underlying general solving mechanism will find an instantiation of variables satisfying the constraints.

Constraint Programming is a powerful problem-solving technology but it is also hard to use; it requires deeper knowledge of how constraint satisfaction works and how the problem should be modelled to be solvable efficiently. The tutorial focuses on these two aspects of CP. First, the tutorial will clarify the traditional misunderstanding that CP is equivalent to enumeration and it will present CP as a technology integrating search and inference. Second, the tutorial will explain why different models of the same problem may lead to dramatically different performance and it will give the audience some guidelines how to design efficiently solvable models.

The tutorial is targeted to a broad research and development community, in particular to those who solve problems of combinatorial nature (scheduling, configuration, assignment, routing etc.). The audience will take away a basic understanding of how constraints work, tips and tricks for efficient constraint modelling, and practical information about writing constraint-based programs.

No specific prior knowledge is required beyond the basic understanding of logic and algorithms.

Roman BartákRoman Barták is a professor at Charles University, Prague (Czech Republic). He leads Constraint Satisfaction and Optimization Research Group that performs basic and applied research in the areas of satisfiability and discrete optimization problems. His work focuses on techniques of constraint satisfaction and their application to planning and scheduling. The research results are used in products of IBM/ILOG and Entellexi. Prof. Barták is teaching courses on artificial intelligence, planning, scheduling, and constraint programming at Charles University and he is author of On-line Guide to Constraint Programming.

T2: Temporal Representation and Reasoning in Interval Temporal Logics
(Half-day, 9:00–12:30)

Angelo Montanari, Pietro Sala and Dario Della Monica


The tutorial aims at giving an account on research in interval temporal logics, with a particular attention to the fragments of Halpern and Shoham's modal logic of time intervals. Interval-based temporal reasoning naturally arises in a variety of AI areas, including knowledge representation and reasoning, temporal planning and maintenance, temporal constraints, theories of actions and events, and event processing. Moreover, it plays a significant role in other fields of computer science, such as theoretical computer science, temporal databases, and computational linguistics.

Our goal is to give an up-to-date and comprehensive description of the state of the art in the field of interval temporal logics. We will illustrate methods and techniques that have been used to deal with the most relevant problems, namely, expressiveness, decidability and complexity of the satisfiability and model checking problems, and reasoning techniques. A special emphasis will be given to the tableau-based methods for interval temporal reasoning. We will conclude the tutorial with a short list of open problems and current research directions.

The natural target of the tutorial are PhD students and researchers who are interested in logical approaches to AI problems. In addition, it is definitely of interest to people dealing with temporal representation and reasoning issues in classical AI application domains, such as, for instance, planning, synthesis, and model-based reasoning.

There are no specific prerequisites, apart from some familiarity with the basics of (propositional) logic and complexity theory. Background knowledge about modal and (point-based) temporal logics would help, but it is not necessary.

Angelo Montanari Angelo Montanari is full professor of Computer Science at the University of Udine. He got his PhD at the University of Amsterdam with a dissertation on logics for time granularity. His research interests are in temporal logics, formal verification, automata theory, game theory, knowledge representation and reasoning, spatial and temporal databases. He is the author of over 170 peer-reviewed publications. He is/was involved in many national and international projects, and in the program committee of many AI and computer science international conferences. He supervised 9 PhD students.

Pietro Sala Pietro Sala is a post-doc at the University of Verona. He got his PhD in Computer Science at the University of Udine. His PhD thesis, entitled: "Decidability of Interval Temporal Logics'", won a prize for the Best Theoretical Computer Science Dissertation awarded by the Italian Chapter of the EATCS. His research interests are in temporal logic, temporal databases, and temporal data mining. He is co-author of over 30 peer-reviewed publications. He taught various courses at the Universities of Udine, Verona, and Trieste.

Dario Della Monica Dario Della Monica got his PhD in Computer Science at the University of Udine. His PhD thesis, entitled: "Expressiveness, decidability, and undecidability of Interval Temporal Logic", won a prize for the Best Dissertation on Computational Logics awarded by the Italian Association for Logic Programming. He has been a post-doc at the University of Salerno, and he is currently a post-doc at Reykjavik University. His research focuses on logic and its applications to computer science (temporal logics, multi-agent systems). He is co-author of over 20 peer-reviewed publications.

T3: The Yesterday, Today, and Tomorrow of Parallelism in Declarative Programming
(Spotlight, 9:00–10:30)

Enrico Pontelli


Since their foundations, declarative programming paradigms have hold the promise of facilitating the exploitation of the computational power of parallel machines. This promise originates from the lack of a rigid control flow structure in declarative programs and the common reliance on mathematical variables (single-assignment). Both these features appear to reduce the presence of sequential dependencies, more easily exposing concurrency and potential parallelism. Nevertheless, this promise has been only partially met over the years, demnstrating that the flexibility of declarative programming in terms of exploitation of parallelism requires solving non-trivial research challenges.

In this tutorial, we will focus on one declarative paradigm (logic programming) and navigate through the history of parallel logic programming, from its inception to present days. The main goals are to recognize the challenges to be resolved to create effective parallel logic programming systems; we will analyze successes, pitfalls, failures, and important lessons learned. We will also characterize the current and most promising lines of research. The tutorial will be organized in a traditional lecture format, with a few breaks among components to allow the audience to ask questions and share experiences.

The tutorial will be accessible to anyone with a minimal understanding of logic programming and its execution models.

Enrico Pontelli Enrico Pontelli is a Regents Professor and Head in the Department of Computer Science at New Mexico State University. His areas of research interest include logic and constraint programming, reasoning about actions and change, bioinformatics, assistive technologies, and computer science education. His research has been supported by several funding agencies, including the National Science Foundation and the Department of Education. He is the recipient of a NSF Career award and a Manasse scholar, and author of over 250 peer-reviewed publications.

T4: Metalearning & Algorithm Selection
(Spotlight, 11:00–12:30)

Pavel Brazdil, Carlos Soares and Joaquin Vanschoren


Algorithm Selection and configuration are increasingly relevant today. Researchers and practitioners from all branches of science and technology face a large choice of parameterized algorithms, with little guidance as to which techniques to use. Moreover, data mining challenges frequently remind us that algorithm selection and configuration are crucial in order to achieve the best performance and drive industrial applications.

Meta-learning leverages knowledge of past algorithm applications to select the best techniques for future applications, and offers effective techniques that are superior to humans both in terms of the end result and especially in a limited time.

In this tutorial, we elucidate the nature of algorithm selection and how it arises in many diverse domains, such as machine learning, data mining, optimization and SAT solving. We show that it is possible to use meta-learning techniques to identify the potentially best algorithm(s) for a new task, based on meta-level information and prior experiments. We also discuss the prerequisites for effective meta-learning systems, and how recent infrastructures, such as OpenML.org, allow us to build systems that effectively advice users on which algorithms to apply.

The intended audience includes researchers (Ph.D.'s), research students and practitioners interested to learn about, or consolidate their knowledge about:

  1. the state-of-the-art in algorithm selection and algorithm configuration;
  2. how to use Data Mining software and platforms to select algorithms in practice;
  3. how to provide advice to end users about which algorithms to select in diverse domains, including optimization, SAT etc. and incorporate this knowledge in new platforms.

The participants should bring their own laptops.

Pavel Brazdil Pavel Brazdil obtained his Ph.D. in AI at the Univ. Edinburgh in 1981 and since then has been lecturing at the Univ. Porto, Portugal (since 1998 as Full Professor). He is a member of LIAAD-INESC Tec Laboratory. His interests lie in data mining, machine learning, meta-learning and text mining. He has supervised 11 Ph.D. students, published/edited 7 books, various chapters in books, journals articles and 40+ articles listed on ISI/DBLP which have achieved 2000+ citations on Google Scholar. He has chaired/organized 10 conferences/workshops (e.g. ECML-91, ECML-93, ECML/PKDD 2005 etc.). He is a Fellow of ECCAI.

Carlos Soares Carlos Soares received a Ph.D. in Computer Science from U.Porto. He works as Associate Professor at the Faculty of Engineering of U.Porto. He is a researcher at INESC TEC, focusing on Data Mining and Business Intelligence. He has participated in 20+ R&ID projects, published/edited several books and 40+ papers in journals and conferences proceedings indexed by ISI. He has participated in the organization of e.g., KDD-09 and ECML-PKDD-12, and he will be Programme Chair for ECML-PKDD-15. He was awarded the Scientific Merit and Excellence Award of the Portuguese AI Association.

Joaquin Vanschoren Joaquin Vanschoren is assistant professor at the Eindhoven University of Technology (TU/e). He runs OpenML.org, an open science platform for machine learning and meta-learning integrated in WEKA, R, and other popular machine learning environments. His research interests include meta-learning, web-scale machine learning, and data science. He obtained several demo and application awards and has been invited speaker on several conferences. He co-organized ECMLPKDD 2013, BeneLearn 2010-2011, as well as the "Silver" (ECMLPKDD) and "Planning to learn" (ECAI) workshops.

AUGUST 18, Afternoon

T5: Software Engineering for Search and Optimization Problems
(Half-day, 14:00–17:30)

Luca Di Gaspero and Tommaso Urli


Over the last years several general purpose open-source and commercial software tools for Search and Optimization have been proposed. Besides the programming language, they mainly differ in terms of the modeling paradigm they support (e.g., Constraint Programming, Local Search, Evolutionary Algorithms, ...) and in their software interface (e.g., Framework, Library, DSL). Moreover, they exhibit different degrees of system integration capabilities and licensing policies, making them adequate just for research prototypes or suitable for industrial projects.

Although a common feeling is that software engineering approaches might be too demanding for research-related software, we advocate that a principled way to organize software design could have a substantial impact also in this field, reducing the development effort in the long run, and fostering positive research behavior, such as the explicit analysis/experimentation of different design choices, or ensuring the reproducibility of computations and results.

In this tutorial we will present a general overview of Search and Optimization tools from a software engineering perspective, highlighting the similarities and differences in the models employed and their impact in the design choices. Moreover, we will show some case studies of a principled development process in different paradigms using the various software tools.

The tutorial is meant for students as well as for researchers and practitioners who may be interested in a practical introduction to software tools and software engineering techniques for Search and Optimization, including researchers interested in understanding the applications of well engineered software tools to the solution of problems in their area.

No specific prerequisite knowledge is required. Some general knowledge of software engineering and object-oriented programming is advisable.

Luca Di Gaspero Luca Di Gaspero is Senior Lecturer of Information Technologies at the University of Udine, Italy. He is member of the SaTT research group, exploring the application of meta-heuristics to real-world problems, mainly in the area of Scheduling, Timetabling, and Logistics. A major focus of his research is in boosting meta-heuristic techniques by their hybridization with AI methods such as Constraint Programming and Reinforcement Learning. He received his MSc (1998) and his PhD (2002) from the University of Udine, Italy. In 2011 he has been a visiting professor at the Vienna University of Technology, Austria.

Tommaso Urli Tommaso Urli is a research fellow and Ph.D. candidate in Computer Science and Optimization at the Department of Electrical, Management and Mechanical Engineering of the University of Udine, Italy. He has been a visiting student at the University of Adelaide, Australia. His main research interests are hybrid meta-heuristics for combinatorial optimization, machine learning, advanced parameter tuning techniques, and problem modeling with exact and heuristic methods. In particular, he is interested in solving sustainable mobility and scheduling problems arising in real-world domains.

T6: Formal Methods for Event Processing
(Half-day, 14:00–17:30)

Alexander Artikis and George Paliouras


Organisations require techniques for automated transformation of the Big Data they collect into operational knowledge. This requirement may be addressed by employing event processing systems that detect activities/events of special significance within an organisation, given streams of low-level information that are very difficult to be utilised by humans.

We will review a Chronicle Recognition System (CRS), the Event Calculus (EC), ProbLog and Markov Logic Networks (MLN). CRS is a purely temporal reasoning system that allows for efficient event processing. EC allows for the representation of temporal and atemporal constraints. Consequently, EC may be used in applications requiring spatial reasoning, for example. ProbLog and MLN, unlike EC and CRS, allow for uncertainty representation and are thus suitable for event processing in noisy environments.

The manual development of event structures is a tedious, time-consuming and error-prone process. Moreover, it is often necessary to update such structures during the event recognition process, due to new information about the application under consideration. For this reason, we will review machine learning techniques automating the construction and refinement of event definitions.

To illustrate the reviewed approaches we will use real-world case studies from the FP7 SPEEDD project: event processing for city transport and traffic management, and credit card fraud management.

The intended audience of the tutorial consists of academics, students and practitioners investigating the open issues of event processing, and/or willing to apply event processing techniques for extracting knowledge from structured and unstructured datasets. Familiarity with AI techniques is desirable.

Alexander Artikis Alexander Artikis is a research associate in NCSR Demokritos (Athens, Greece). He holds a PhD from Imperial College London on norm-governed multi-agent systems, while his research interests lie in the areas of AI and distributed systems. Alexander has been working on several international projects on event processing; currently, he is the technical director of the FP7 SPEEDD project that develops a Big Data system for proactive event-driven decision-making. Alexander has also developed a highly scalable, logic-based, open-source event processing system.

George Paliouras Georgios Paliouras is a senior researcher in NCSR Demokritos (Athens, Greece). He holds a PhD from Manchester University on machine learning for event recognition. He has performed basic and applied research in machine learning for the last 17 years. He is involved in many European and national research projects and has the role of scientific coordinator in some of them, including the FP7 SPEEDD project. He has given a number of invited talks and tutorials at various institutions and conferences, such as ECML/PKDD, DEBS and IJCAI.

T7: Tabling for Planning
(Spotlight, 14:00–15:30)

Neng-Fa Zhou


This tutorial introduces a tabling approach to planning. In particular, it describes the planner module of the Picat language and its implementation. It also introduces the modeling methodology through several planning examples.

Picat is a Prolog-like multi-paradigm programming language that provides pattern matching, deterministic and non-deterministic rules, loops, functions, constraints, and tabling as its core modeling and solving features. The planner module uses tabling. For a planning problem, users only need to specify conditions on final states and a set of actions, and call the planner on an initial state to find a plan or a best plan.

The Picat planner treats a planning problem as a state-space search problem. It tables all the states that have been encountered during search. Tabling prevents looping, and term-sharing techniques alleviate the state explosion problem. The tabling system is tailored to planning with two new techniques, called early termination and resource-bounded search. The planner returns a satisfactory plan as an exception without waiting for the completion of the tabled predicate. In resource-bounded search, the planner fails a state immediately if the same state has failed before and the new occurrence does not carry more resources than the old one.

This tutorial will present several modeling examples and show the competitiveness of tabled planning in comparison with the cutting-edge ASP and PDDL planners.

The tutorial is targeted to researchers in the areas of planning, logic programming, and answer set programming. No special prerequisites are expected, though knowledge of logic programming is an advantage.

Neng-Fa Zhou Neng-Fa Zhou is a professor at the City University of New York (Brooklyn College and Graduate Center). He has published over 40 papers in major journals and conferences on programming languages, constraint-solving, planning, graphics, and machine learning systems. He is the main designer and implementer of B-Prolog and Picat. He has also participated in the development of PRISM, which runs on top of B-Prolog. He has been an active participant in various competitions, including CSP, ASP, Minizinc, and Prolog Programming Contest.

T8: Probabilistic Programming
(Spotlight, 16:00–17:30)

Luc De Raedt and Angelika Kimmig


Probabilistic programming is an emerging subfield of AI that extends traditional programming languages with primitives to support probabilistic inference and learning. It is closely related to statistical relational learning, but focuses on a programming language perspective rather than on a graphical model one.

This tutorial provides a gentle and coherent introduction to the field by introducing a number of core probabilistic programming concepts and their relations. It focuses on probabilistic extensions of logic programming languages, such as CLP(BN), BLPs, ICL, PRISM, ProbLog, LPADs, CP-logic, SLPs and DYNA, but also discusses relations to alternative probabilistic programming languages such as Church, IBAL and BLOG and to some extent to statistical relational learning models such as RBNs, MLNs, and PRMs.

The concepts will be illustrated on a wide variety of tasks, including models representing Bayesian networks, probabilistic graphs, stochastic grammars, etc. This should allow participants to start writing their own probabilistic programs. We further provide an overview of the different inference mechanisms developed in the field, and discuss their suitability for the different concepts. We also touch upon approaches to learn the parameters of probabilistic programs, and mention a number of applications in areas such as robotics, vision, natural language processing, web mining, and bioinformatics.

The tutorial is intended for AI researchers and practitioners, as well as domain experts interested in probabilistic programming and statistical relational learning. Basic knowledge of Prolog, logic programming and/or graphical models at the level of an introductory course in AI will be helpful, but is not a prerequisite.

Luc De Raedt Luc De Raedt is a full professor at the KU Leuven, where he also obtained his Ph.D. in 1991. From 1999-2006 he was a full professor at the Albert-Ludwigs-University Freiburg. His research interest focusses on statistical relational learning and combining constraint programming with machine learning. He is also interested in applications in bioinformatics, robotics, vision and natural language processing. He was a program-chair of conferences such as ECML/PKDD 2001, ICML 2005 and ECAI 2014 and coordinated a number of European projects such as ILP, APRIL and ICON.

Angelika Kimmig Angelika Kimmig is a postdoctoral fellow of the Research Foundation - Flanders (FWO Vlaanderen) at KU Leuven, Belgium, where she obtained here Ph.D. in 2010, and has recently spent a year at the University of Maryland, College Park, USA. Her research interests include symbolic AI, reasoning under uncertainty, machine learning, logic programming, and especially combinations thereof such as probabilistic programming and statistical relational learning. Angelika is one of the key developers of the probabilistic logic programming language ProbLog.

AUGUST 19, Morning

T9: Introduction to Statistical and Behavioural Analysis of Agent-Based Models
(Half-day, 9:00–12:30)

Tamás Máhr, László Gulyás and Richárd O. Legéndi


Multi-agent simulations are gaining ground in many disciplines, in particular in the social sciences. These simulations often consist of rather simple agents, focusing more on the emergent society-level behavior, rather than on individual cognitive abilities. Nonetheless, the method itself does not exclude the modeling of cognitively complex agents. In any case, a methodologically correct exploration of the behavior of the models created is of paramount importance for the creation of robust, reproducible results. In addition, techniques that help the discovery of actual human behavior may also help developing more realistic artificial agents. In our tutorial, we will offer and discuss tools to help researchers with these tasks.

We will introduce two frameworks intended to help modellers to explore the behaviour of their models either by performing smart parameter space explorations and/or participatory experiments by executing their models within a web-based environment:

  1. MEME, a generic tool that supports model analysis over RepastJ, NetLogo, and Mason, designed to run large-scale parameter space explorations on grid/cloud systems or sensitivity analysis through statistical methods;
  2. PET v2.0, a web application that incorporates agent-based simulations into a web interface enabling users to administrate, run and participate in simulations

The tutorial has no particular requirements, but experience in implementing agent-based models is an advantage. The frameworks are introduced through a supplied model, no programming is necessary.

Tamás Máhr Tamás Máhr is project manager and senior developer at the simulation and intelligent applications group of AITIA International Inc. He obtained his PhD from the Delft University of Technology in the topic of robustness of multi-agent logistical planning. Since working at AITIA, he has been managing EC collaborative projects (FP7), and been active in the development of MASS. He focuses on tools that facilitate the exploration of agent-based models by allowing researchers to apply predefined experimental designs in computational simulations. Furthermore, he contributes to a JavaEE-based tool that turns Mason simulations into web applications.

László Gulyás László Gulyás is assistant professor at the Department of History and Philosophy of Science, Eotvos Lorand University, Budapest, and a research partner at AITIA International, Inc, where he leads the development of the Multi-Agent Simulation Suite (MASS). He's been teaching agent-based modeling and simulation at Harvard University and at the Central-European University. He co-directed the Complex Systems and Social Simulation summer school at the Central European University's Summer University, Budapest, 2009-2010, and was also a faculty member at the 2002 Budapest Complex Systems Summer School organized by the Santa Fe Institute.

Richárd O. Legéndi Richárd Olivér Legéndi is a Ph.D. student at Eotvos Lorand University, Budapest, Hungary. His research interests include agent-based modeling and dynamic networks, mainly focusing on various aspects and applications of Agent-Based Simulations. His Ph.D. research focuses on dynamic information and social networks. Problems he investigate are motivated by the need to understand the inherent properties brought by the evolution of networks and changes over time. He is a developer of the Multi-Agent Simulation Suite (MASS) and one of the main developers of the Functional Agent-Based Language for Simulations (FABLES).

T10: Procedural Content Generation in Games
(Spotlight, 9:00–10:30)

Noor Shaker


Procedural Content Generation (PCG), as defined as the automatic generation of game content with limited or no input from human designer/player, is a rapidly growing field offering hope for substantially reducing the authoring burden in games, improving our theoretical and empirical understanding of game design and its influence on players, and enabling entirely new kinds of games and playable experiences.

The tutorial will present an overview of the different methods of PCG and its application to several game genres. This will be supported by several illustrative examples from research and the industry. The tutorial will cover key techniques employed by PCG such as search-based approaches, constructive methods and answer set programming and their applications for creating different types of content such as maps, levels, weapons, tracks, game rules and even complete games. The tutorial will also discuss the Player-Driven PCG paradigm that focuses on generating player-adapted content and the mix-initiative approach that aims at incorporating human designer/player input with computer- generated content.

This tutorial is intended to bridge the gap between researchers who are working on applying AI techniques in the PCG framework, and those working on advancing the state-of-the-art of AI and its applications to other problems. The tutorial is expected to attract researchers with little or no previous experience or knowledge about PCG, but who are interested in learning about it.

As prerequisites basic knowledge in AI search methods and propositional logic is assumed.

Noor Shaker Noor Shaker is a Postdoctoral researcher at the IT University of Copenhagen. She received a 5-year BA in IT Engineering in 2007 from Damascus University, an M.Sc. in Artificial Intelligence in 2009 from Katholieke Universiteit Leuven and a Ph.D. degree from the IT University of Copenhagen in 2013. Her research interests include player modeling, procedural content generation and human-computer interaction. She is the chair of IEEE CIS Task Force on Player Modeling, and a Technical Committee Member of the IEEE Computational Intelligence and Games Society.

T11: Replication and Recomputation in Scientific Experiments
(Spotlight, 11:00–12:30)

Ian Gent and Lars Kotthoff


This tutorial will discuss and demonstrate the general principles of recomputation. By recomputation, we mean the replication of experiments in computational science.

The tutorial will be in two parts. First we will discuss the nature of computational experiments and why Computer Science would be better if experiments were made available for recomputation as standard practice. This will be based on The Recomputation Manifesto by Ian Gent, which argues that recomputation should become a standard part of Computer Science experimental methodology. The second part of the tutorial will discuss methods for making experiments recomputable.

The intended audience is any scientist in Artificial Intelligence who engages in computational experiments. This means almost any practitioner of AI except pure theoreticians.

We will invite participants in the tutorial to submit experiments for past and future papers to an experimental repository at recomputation.org.

Ian Gent Ian Gent is Professor of Computer Science at the University of St Andrews, Scotland. His interest in the proper foundations of empirical science in computing date almost 20 years. He has given tutorials on “Empirical Methods in CS and AI” at conferences such as IJCAI 2001. Of his non peer-reviewed papers, his most cited by far is “How Not To Do It”, a collection of embarrassing mistakes he and colleagues have made in computational experiments. To show how good we are at not doing things right, we mis-spelt the name of one of the authors!

Lars Kotthoff Lars Kotthoff is Postdoctoral Research Fellow at University College Cork. His research work has often involved extensive experimentation. He has held funded research grants from Amazon and Rackspace on investigating the use of clouds for scientific experimentation.

AUGUST 19, Afternoon

T12: Search Methods for Classical and Temporal Planning
(Half-day, 14:00–17:30)

Jussi Rintanen


During the last ten years a number of algorithmic breakthroughs have lifted the efficiency of planning methods to the level required in many real-world applications. The tutorial presents the most important approaches to solving reachability problems in untimed (asynchronous) and timed transition systems, as applied in algorithms for classical/deterministic planning and for temporal planning. These include planning as satisfiability (including SAT modulo Theories), planning by heuristic state-space search, and logic-based symbolic techniques for state space traversal.

As prerequisites basic knowledge in AI search methods and propositional logic is assumed.

Jussi Rintanen Jussi Rintanen got his doctoral degree from Helsinki University of Technology in 1997, and has since worked at the Universities of Ulm, Albert-Ludwigs University Freiburg (Germany), National ICT Australia, and since 2012 at Aalto University (former Helsinki University of Technology) in Helsinki, Finland. His research focuses on the use of combinatorial search methods, automated reasoning, and applications of logics in computer science, including plan and program synthesis, diagnosis and other modes of reasoning.

T13: Intelligent Socio Technical Interaction
(Spotlight, 14:00–15:30)

Virginia Dignum and Frank Dignum


Increasingly, software systems mediate between people or function as a partner in accomplishing joint tasks. As intermediaries between people or as partners in accomplishing joint tasks, software systems get more responsibility to start tasks, to react to failures and to find alternative ways to accomplish a task, that is, to act as if they are members of a team, rather than being passive tools. This shift requires ICT-applications to become aware of and act socially intelligent in their social situation.

The overall aim of this tutorial is to present new computational theories of social intelligence that are needed to create intelligent socio-technical interactions. During the tutorial we will discuss how these theories can be used for the implementation of socially intelligent systems. E.g. an agent might agree to provide help, because it wants to be a good friend and providing help is expected in that situation (even though the agent might not know how to provide the help yet).

We will cover the following topics:

  1. Reasoning about social aspects.
  2. Setting social goals.
  3. Reasoning about the social environment and action.
  4. Recognizing, reasoning about, and using the environment for social aims.
  5. Incorporating social motivations and attitudes with achievement goals.

The intended audience of this tutorial are researchers in social aspects of intelligent behavior, intelligent virtual agents, agent based social simulation, intelligent interaction or dialogue management.

Both experienced as well as novice researchers can benefit from this tutorial. There is no required prior knowledge.

Virginia Dignum Virginia Dignum (1964) is Associate professor at the Faculty of Technology, Policy and Management, Delft University of Technology. She got her PhD in 2004 at Utrecht University. Her research focuses on agent based models of organizations, and the interaction between people and intelligent systems and teams. In 2006, she was awarded the prestigious Veni grant from NWO (Dutch Organization for Scientific Research). She has organized many international conferences and workshops, including AAMAS2005 and ECAI2016, and has more than 150 peer-reviewed publications and books.

Frank Dignum Frank Dignum (1961) is Associate professor at Information and Computing Sciences department of Utrecht University. Frank Dignum received a PhD. from the Free University of Amsterdam, The Netherlands in 1989. He has published more than 200 papers and several books. He has also organized several conferences and workshops amongst which AAMAS2005 and ECAI2016. He was keynote and invited speaker and given many tutorials and seminars around the world. His main research foci are on social aspects of agents, agents and services, agents for games and agent based social simulation.

T14: Multilingual Semantic Processing with BabelNet
(Spotlight, 16:00–17:30)

Roberto Navigli and Andrea Moro


We introduce concepts of Multilingual Semantic Processing by first surveying important challenges in multilingual semantic processing and then introducing solutions to those challenges which use BabelNet, a freely-available multilingual encyclopaedic dictionary and semantic network.

The tutorial will be organized as follows:

  1. Foundations: Motivation and background in semantic processing for understanding the tutorial content.
  2. Constructing a multilingual semantic resource: Challenges in building a common multilingual representation. We describe a case study from BabelNet which merges two common semantic resources, Wikipedia and WordNet, and we then introduce recent efforts to merge other collaboratively-constructed semantic resources such as OmegaWiki and Wiktionary. We provide insight into issues such as aligning semantic representations across multiple languages and resources.
  3. Multilingual Word Sense Disambiguation and Entity Linking: Challenges of representing ambiguous, multilingual text as specific concepts and named entities.
  4. Adding multilingual semantics to Information Extraction: Techniques for multilingual relation extraction in an (open) Information Extraction setting.

We will provide references, code examples, and exercises that participants can further explore after the tutorial session is over.

The audience should consist of PhD students, researchers, or members of industry who have interest in going from the monolingual to the multilingual setting and increasing their understanding of using multilingual data/representations. We do not assume that the audience has any experience with ontologies or is familiar with multiple languages.

Roberto Navigli Roberto Navigli is an associate professor in the Department of Computer Science at the Sapienza University of Rome. He is the recipient of an ERC Starting Grant in computer science and informatics on multilingual word sense disambiguation (2011-2016) and a co-PI of a Google Focused Research Award on Natural Language Understanding. His research interests lie in the field of Natural Language Processing.

Andrea Moro Andrea Moro is a PhD student in the Department of Computer Science at the Sapienza University of Rome. His research interests focus on Natural Language Understanding with an emphasis on Unsupervised Relation Extraction, Word Sense Disambiguation and Entity Linking.



ECAI Tutorial chairs

Agostino Dovier   Paolo Torroni
Agostino Dovier Paolo Torroni