Program
Monday, February 17, 2020
09:00 - 09:15: Opening Session |
09:15 - 10:15: Keynote Talk Jan Van den Bussche: First-order logic of information flows: expressibility, primitivity, input-output properties, and limited access patterns |
10:15 - 10:45: Coffee Break |
10:45 - 12:15: Session "Knowledge Representation and Automated Reasoning 1" Stipe Pandzic: On the dynamics of structured argumentation: Modeling changes in default justification logic Wolfgang Dvorák, Anna Rapberger, Johannes P. Wallner and Stefan Woltran: ASPARTIX-V19 - An Answer-set Programming based System for Abstract Argumentation (short talk) |
12:15 - 13:45: Lunch Break |
13:45 - 15:15: Session "Distributed and Parallel Databases, Distributed Knowledge 1" Klaus-Dieter Schewe, Yamine Ait Ameur and Sarah Benyagoub: Realisability of Choreographies Daniel Xia, Michael Simpson, Venkatesh Srinivasan and Alex Thomo: Strongly Minimal MapReduce Algorithms: A TeraSort Case Study (short talk) |
15:15 - 15:45: Coffee Break |
15:45 - 17:45: Session "Databases, Dependencies, and Dependence Logic 1" Robert Brijder, Marc Gyssens and Jan Van den Bussche: On matrices and K-relations Munqath Al-Atar and Attila Sali: Functional Dependencies in Incomplete Databases With Limited Domains (short talk) Bernhard Thalheim: Schema Optimisation Instead Of (Local) Normalisation (short talk) |
Tuesday, February 18, 2020
09:00 - 10:00: Keynote Talk Bernhard Nebel: Implicit coordination of mobile agents |
10:00 - 10:30: Coffee Break |
10:30 - 12:30: Session "Databases, Dependencies, and Dependence Logic 2" Yasir Mahmood and Arne Meier: Parameterised Complexity of Model Checking and Satisfiability in Propositional Dependence Logic Pietro Galliani and Jouko Vaananen: Diversity, dependence and independence |
12:30 - 14:00: Lunch Break |
14:00 - 15:30: Session "Interaction between Multiple Agents" Yuri David Santos: Social Consolidations: Rational Belief in a Many-Valued Logic of Evidence and Peerhood Munyque Mittelmann and Laurent Perrussel: Game Description Logic with Integers: A GDL Numerical Extension (short talk) |
15:30 - 16:00: Coffee Break |
16:00 - 18:00: Session "Knowledge Representation and Automated Reasoning 2" Silvia Ghilezan, Jelena Ivetic, Simona Kasterovic, Zoran Ognjanovic and Nenad Savic: Towards probabilistic reasoning in type theory - the intersection type case Christoph Beierle and Jonas Haldimann: Normal Forms of Conditional Knowledge Bases Respecting Entailments and Renamings |
18:00: FoIKS Steering Committee Meeting (open to everyone) |
Wednesday, February 19, 2020
09:00 - 10:00: Keynote Talk (EurAI Talk) Anni-Yasmin Turhan: Description Logics for Typicality and Nonmonotonic Reasoning |
10:00 - 10:30: Coffee Break |
10:30 - 12:30: Session "Ontologies and Applications" Xianglin Zhan, Cai Lu and Guangmin Hu: Event Sequence Interpretation of Structural Geomodels: A Knowledge-Based Approach for Extracting Tectonic Sequences Dennis Medved, Johan Nilsson and Pierre Nugues: Utilizing Deep Learning and RDF to Predict Heart Transplantation Survival |
12:30 - 14:00: Lunch Break |
14:00: Social Event and Conference Dinner |
Thursday, February 20, 2020
09:00 - 10:00: Keynote Talk Klaus-Dieter Schewe: Computation on structures - behavioural theory, logic and complexity |
10:00 - 10:30: Coffee Break |
10:30 - 12:30: Session "Distributed and Parallel Databases, Distributed Knowledge 2" Elena Ravve: Logic-based Approach to Incremental Monitoring and Optimization on Strongly Distributed Data Streams Ryo Murai and Katsuhiko Sano: Craig Interpolation of Epistemic Logics with Distributed Knowledge |
12:30 - 14:00: Lunch Break |
14:00 - 15:30: Session "Inconsistency and Complexity" John Grant and Francesco Parisi: Measuring Inconsistency in a General Information Space Flavio Ferrarotti, Senén González, Klaus-Dieter Schewe and Jose Turull-Torres: Proper Hierarchies in Polylogarithmic Time and Absence of Complete Problems (short talk) |
15:30 - 16:00: Closing and Coffee |
Social Event and Conference Dinner
The social event takes us to the Zollern Colliery which is located in the west of Dortmund. It is one of the most beautiful and impressive testimonies to Germany's industrial history. It was closed down in 1966 and serves as the headquarters of the Westphalian Industrial Museum now. We join a guided tour in English at 3:30 p.m. but there is also some time to inspect the museum on your own.
The conference dinner takes place at HÖVELS Hausbrauerei in the city of Dortmund. We have arranged
a tour of the HÖVELS in-house brewery at 6 p.m. right before the dinner, too.
Accepted Papers
- Wolfgang Dvořák, Anna Rapberger, Johannes P. Wallner and Stefan Woltran: ASPARTIX-V19 - An Answer-set Programming based System for Abstract Argumentation
- Robert Brijder, Marc Gyssens and Jan Van den Bussche: On matrices and K-relations
- Stipe Pandžić: On the dynamics of structured argumentation: Modeling changes in default justification logic
- Pietro Galliani and Jouko Vaananen: Diversity, dependence and independence
- Ryo Murai and Katsuhiko Sano: Craig Interpolation of Epistemic Logics with Distributed Knowledge
- John Grant and Francesco Parisi: Measuring Inconsistency in a General Information Space
- Klaus-Dieter Schewe, Yamine Ait Ameur and Sarah Benyagoub: Realisability of Choreographies
- Yasir Mahmood and Arne Meier: Parameterised Complexity of Model Checking and Satisfiability in Propositional Dependence Logic
- Elena Ravve: Logic-based Approach to Incremental Monitoring and Optimization on Strongly Distributed Data Streams
- Christoph Beierle and Jonas Haldimann: Normal Forms of Conditional Knowledge Bases Respecting Entailments and Renamings
- Munyque Mittelmann and Laurent Perrussel: Game Description Logic with Integers: A GDL Numerical Extension
- Yuri David Santos: Social Consolidations: Rational Belief in a Many-Valued Logic of Evidence and Peerhood
- Dennis Medved, Johan Nilsson and Pierre Nugues: Utilizing Deep Learning and RDF to Predict Heart Transplantation Survival
- Silvia Ghilezan, Jelena Ivetić, Simona Kašterović, Zoran Ognjanović and Nenad Savić: Towards probabilistic reasoning in type theory - the intersection type case
- Xianglin Zhan, Cai Lu and Guangmin Hu: Event Sequence Interpretation of Structural Geomodels: A Knowledge-Based Approach for Extracting Tectonic Sequences
- Bernhard Thalheim: Schema Optimisation Instead Of (Local) Normalisation
- Munqath Al-Atar and Attila Sali: Functional Dependencies in Incomplete Databases With Limited Domains
- Flavio Ferrarotti, Senén González, Klaus-Dieter Schewe and Jose Turull-Torres: Proper Hierarchies in Polylogarithmic Time and Absence of Complete Problems
- Daniel Xia, Michael Simpson, Venkatesh Srinivasan and Alex Thomo: Strongly Minimal MapReduce Algorithms: A TeraSort Case Study
Keynotes
Jan Van den Bussche
Jan Van den Bussche is professor of databases and theoretical
computer science at Hasselt University in Belgium. He received
his PhD from the University of Antwerp in 1993, under the
supervision of Jan Paredaens. He served as PC chair, and chair
of the council, for the International Conference on Database
Theory, and also as PC chair, and chair of the executive
committee, for the ACM Symposium on Principles of Database
Systems. His main research interest is in data models and query
languages for a wide variety of data applications, ranging from
spatial data to data stored in DNA. Most recently he is leading
a work package on distributed data intelligence within the
context of the Artificial Intelligence Research Flanders
initiative.
First-order logic of information flows: expressibility,
primitivity, input-output properties, and limited access patterns
The logic of information flows has been proposed by Ternovska as a general framework in the field of knowledge representation. The general aim of LIF is to model how information propagates in complex systems. In this work, we focus on the first-order version of LIF. Unique to LIF is the dynamic semantics which relates input assignments to output assignments. We formulate semantic definitions of input and output variables of a LIF expression. These semantic definitions are then approximated by syntactic definitions. The expressive power of LIF is compared to that of first-order logic (FO), and bounded-variable fragments of FO. In this setting, we investigate the primitivity of the composition operator in LIF. We also introduce the "forward" fragment of LIF and show how it can offer a fresh, navigational, perspective on querying information sources with limited access patterns.
The logic of information flows has been proposed by Ternovska as a general framework in the field of knowledge representation. The general aim of LIF is to model how information propagates in complex systems. In this work, we focus on the first-order version of LIF. Unique to LIF is the dynamic semantics which relates input assignments to output assignments. We formulate semantic definitions of input and output variables of a LIF expression. These semantic definitions are then approximated by syntactic definitions. The expressive power of LIF is compared to that of first-order logic (FO), and bounded-variable fragments of FO. In this setting, we investigate the primitivity of the composition operator in LIF. We also introduce the "forward" fragment of LIF and show how it can offer a fresh, navigational, perspective on querying information sources with limited access patterns.
Bernhard Nebel
Bernhard Nebel received his first degree in Computer Science (Dipl.-Inform.) from the University of Hamburg and his Ph.D. (Dr. rer. nat.) from the University of Saarland in 1989. Between 1982 and 1993 he worked on different AI projects at the University of Hamburg, the Technical University of Berlin, ISI/USC, IBM Germany, and the German Research Center for AI (DFKI). From 1993 to 1996 he held an Associate Professor position at the University of Ulm. Since 1996 he is Professor at Albert-Ludwigs-Universität Freiburg and head of the research group on Foundations of Artificial Intelligence. Bernhard Nebel is an EurAI and AAAI fellow and member of the Leopoldina. His research interests are knowledge representation, planning, and the application of methods from these areas in robotic contexts.
Implicit coordination of mobile agents
In multi-agent path finding (MAPF), it is usually assumed that planning is performed centrally and that the destinations of the agents are common knowledge. We will drop both assumptions and analyze under which conditions it can be guaranteed that the agents reach their respective destinations using implicitly coordinated plans without communication. Furthermore, we will analyze what the computational costs associated with such a coordination regime are. As it turns out, guarantees can be given assuming that the agents are of a certain type. However, the implied computational costs are quite severe. The plan existence problem becomes PSPACE-complete, while the original MAPF plan existence problem is in P. This clearly demonstrates the value of communicating about plans before execution starts.
In multi-agent path finding (MAPF), it is usually assumed that planning is performed centrally and that the destinations of the agents are common knowledge. We will drop both assumptions and analyze under which conditions it can be guaranteed that the agents reach their respective destinations using implicitly coordinated plans without communication. Furthermore, we will analyze what the computational costs associated with such a coordination regime are. As it turns out, guarantees can be given assuming that the agents are of a certain type. However, the implied computational costs are quite severe. The plan existence problem becomes PSPACE-complete, while the original MAPF plan existence problem is in P. This clearly demonstrates the value of communicating about plans before execution starts.
Klaus-Dieter Schewe
Klaus-Dieter Schewe studied mathematics and computer
science at the University of Bonn, Germany. Originally he worked in
group representation theory, which was also the field of his Ph.D. After
some years in industry he returned to academia working in software
engineering, knowledge representation, database theory and rigorous
methods, bringing these diverse fields together. His particular interest
is on mathematical and logical foundations, semantics and
expressiveness. He investigated methods for consistency enforcement,
contributed to many results in dependency theory for complex-value
databases (together with Sali, Link and Hartmann), developed a thorough
methodology for the design and development of web information systems
(together with Thalheim), created a client-centric middleware for cloud
computing (together with Bosa and others), and developed a theory of
knowledge patterns for entity resolution (together with Qing Wang). He
developed behavioural theories for unbounded parallel and reflective
algorithms (together with Ferrarotti, Wang and Tec) and for concurrent
systems (together with Börger), with which he contributed to the
foundations of rigorous methods. Recently, he linked this research to
descriptive complexity theory. He graduated in Pure Mathematics at
University of Bonn, received a Ph.D. 1985 from University of Bonn in
Mathematics, and later in 1995 received his D.Sc. from Brandenburg
Technical University in Theoretical Computer Science. He was Chair of
Information Systems at Massey University Director of the Information
Science Research Centre in New Zealand, and Scientific Director of the
Software Competence Center Hagenberg in Austria. In September 2019 he
joined the University of Illinois at Urbana Champaign Institute of
Zhejiang University at the International Campus in Haining, China.
Computation on Structures - Behavioural Theory, Logic and Complexity
Database theory is closely linked to finite model theory. Each relational database defines a finite relational structure, and queries as well as database transformations are computations on these structures. Furthermore, descriptive complexity theory provides many links between the expressiveness of logics concerning queries and common space/time complexity classes. Examples are the celebrated results by Fagin concerning the capture of NP by existential second-order logic and by Immerman/Vardi concerning the capture of PTIME by first-order logic plus inflationary fixed-point over ordered structures. Another related development is provided by behavioural theories of classes of algorithms and proofs that these are captured by variants of abstract state machines, which operate on isomorphism classes of structures. Such theories with purely logical definitions of classes of computations of interest have been developed for sequential, parallel, concurrent and reflective algorithms. Therefore, we will argue for a shift of paradigm making computations on structures as provided by abstract state machines the standard model of computation with the advantage that computations on arbitrary levels of abstraction are enabled. Behavioural theories provide the means for logical characterisation of classes of computations, and also the logical characterisation of complexity in connection with the capture of complexity classes by specific classes of abstract state machines seems possible and adequate. We will provide evidence for the rationale of this shift.
Database theory is closely linked to finite model theory. Each relational database defines a finite relational structure, and queries as well as database transformations are computations on these structures. Furthermore, descriptive complexity theory provides many links between the expressiveness of logics concerning queries and common space/time complexity classes. Examples are the celebrated results by Fagin concerning the capture of NP by existential second-order logic and by Immerman/Vardi concerning the capture of PTIME by first-order logic plus inflationary fixed-point over ordered structures. Another related development is provided by behavioural theories of classes of algorithms and proofs that these are captured by variants of abstract state machines, which operate on isomorphism classes of structures. Such theories with purely logical definitions of classes of computations of interest have been developed for sequential, parallel, concurrent and reflective algorithms. Therefore, we will argue for a shift of paradigm making computations on structures as provided by abstract state machines the standard model of computation with the advantage that computations on arbitrary levels of abstraction are enabled. Behavioural theories provide the means for logical characterisation of classes of computations, and also the logical characterisation of complexity in connection with the capture of complexity classes by specific classes of abstract state machines seems possible and adequate. We will provide evidence for the rationale of this shift.
Anni-Yasmin Turhan
Anni-Yasmin Turhan is an assistant professor at the Institute of Theoretical Computer Science of Dresden University of Technology.
Having obtained her Diploma in Computer Science from the University of Hamburg, she started her scientific career in a research project dedicated to non-standard
inferences in Description Logics (DLs). She received her doctoral (2008) and habilitation degree (2014) from TU Dresden. Her research interests are rooted
in the fields of artificial intelligence and theoretical computer science and are dedicated to knowledge representation and reasoning. She has been a program
committee member for many conferences in the fields of knowledge representation and reasoning, Semantic Web, and AI. In her research she studied inferences
that can be employed to build and maintain DL knowledge bases. Her recent research projects are dedicated to situation recognition by DL reasoning, reasoning
using similarity, and reasoning under nonmonotonic semantics.
Description Logics for Typicality and Nonmonotonic Reasoning
Description Logics (DLs) are an intensively studied class of logics tailored towards building ontologies. Most DLs are decidable fragments of first-order logic. A DL knowledge base consists of two parts: the terminological part that captures the concepts of the application at hand by means of concept axioms and of the assertional part that captures the data from the application. In practical applications where exceptions occur frequently, such as Biology or Medicine, classical, monotone reasoning is often undesirable. Likewise, for applications where data may be erroneous and causes inconsistencies w.r.t. the DL terminology, monotonic reasoning is unsuitable. An approach to achieve nonmonotonic behaviour of DLs is to employ defeasible description logics, which admit the use of defeasible concept axioms. During reasoning these axioms can be omitted for those instances that otherwise would cause an inconsistency. Defeasible DLs can also be employed to model typicality. Instances that fulfill more defeasible axioms are simply regarded more typical than others. Over the last decade a whole range of defeasible DLs has been proposed in the literature. The proposals have very different strengths and weaknesses and can be rather hard to compare with each other - in particular, as the semantics are often given in a procedural way. In this talk I introduce defeasible DLs and demonstrate different approaches to reasoning in these DLs. I discuss our recent approach for the defeasible Description Logic EL_\bot that has model-theoretic semantics and remedies some of the severe shortcomings of earlier approaches. We also investigate the complexity of reasoning in this defeasible DL in different settings.
Description Logics (DLs) are an intensively studied class of logics tailored towards building ontologies. Most DLs are decidable fragments of first-order logic. A DL knowledge base consists of two parts: the terminological part that captures the concepts of the application at hand by means of concept axioms and of the assertional part that captures the data from the application. In practical applications where exceptions occur frequently, such as Biology or Medicine, classical, monotone reasoning is often undesirable. Likewise, for applications where data may be erroneous and causes inconsistencies w.r.t. the DL terminology, monotonic reasoning is unsuitable. An approach to achieve nonmonotonic behaviour of DLs is to employ defeasible description logics, which admit the use of defeasible concept axioms. During reasoning these axioms can be omitted for those instances that otherwise would cause an inconsistency. Defeasible DLs can also be employed to model typicality. Instances that fulfill more defeasible axioms are simply regarded more typical than others. Over the last decade a whole range of defeasible DLs has been proposed in the literature. The proposals have very different strengths and weaknesses and can be rather hard to compare with each other - in particular, as the semantics are often given in a procedural way. In this talk I introduce defeasible DLs and demonstrate different approaches to reasoning in these DLs. I discuss our recent approach for the defeasible Description Logic EL_\bot that has model-theoretic semantics and remedies some of the severe shortcomings of earlier approaches. We also investigate the complexity of reasoning in this defeasible DL in different settings.