ABSTRACT. In order to understand natural language expressions it is usually not enough to know the literal (dictionary) meaning of words used in these expressions and compositional rules of the corresponding language. Much more knowledge is involved here; knowledge, which may have nothing to do with the linguistic competence but is rather related to our general conception of the world. In this tutorial, I will present an overview of the approaches to modeling knowledge-intensive natural language understanding (NLU) in a computational framework, with the main focus on inference-based NLU. I will discuss the types of knowledge required for NLU and the techniques to obtain this knowledge in a machine-readable form. The tutorial will cover lexical- semantic knowledge bases, ontologies (general and domain-specific), and corpora-based resources. Then I will focus on reasoning procedures applicable to NLU and compare two main forms of logical inference: deduction and abduction. In the last part of the tutorial, I will present applications of some of the recent large-scale inference-based NLU systems to such knowledge-intensive tasks as recognizing textual entailment, metaphor interpretation, and narrativization of videos.

ABSTRACT. Dynamic epistemic logic (DEL) takes the classical topic epistemic logic and adds features from propositional dynamic logic (PDL). DEL comes to life in the multiagent setting. It allows one to represent epistemic actions in addition to epistemic situations. DEL has been around for about 20 years, and so it has seen a lot of progress. By now, the original topics have been applied, generalized, and adapted in various ways. My primary goal in this tutorial is to present an overview of the area to people at KR 2014. The talk will discuss recent work in KR that uses DEL. After this, I would like to ask why it is that tools from DEL are not more used in the KR community. Turning things around, I also ask what influence issues from KR and other fields will have on the future of DEL and related frameworks.

ABSTRACT. In order to understand natural language expressions it is usually not enough to know the literal (dictionary) meaning of words used in these expressions and compositional rules of the corresponding language. Much more knowledge is involved here; knowledge, which may have nothing to do with the linguistic competence but is rather related to our general conception of the world. In this tutorial, I will present an overview of the approaches to modeling knowledge-intensive natural language understanding (NLU) in a computational framework, with the main focus on inference-based NLU. I will discuss the types of knowledge required for NLU and the techniques to obtain this knowledge in a machine-readable form. The tutorial will cover lexical- semantic knowledge bases, ontologies (general and domain-specific), and corpora-based resources. Then I will focus on reasoning procedures applicable to NLU and compare two main forms of logical inference: deduction and abduction. In the last part of the tutorial, I will present applications of some of the recent large-scale inference-based NLU systems to such knowledge-intensive tasks as recognizing textual entailment, metaphor interpretation, and narrativization of videos.

ABSTRACT. Dynamic epistemic logic (DEL) takes the classical topic epistemic logic and adds features from propositional dynamic logic (PDL). DEL comes to life in the multiagent setting. It allows one to represent epistemic actions in addition to epistemic situations. DEL has been around for about 20 years, and so it has seen a lot of progress. By now, the original topics have been applied, generalized, and adapted in various ways. My primary goal in this tutorial is to present an overview of the area to people at KR 2014. The talk will discuss recent work in KR that uses DEL. After this, I would like to ask why it is that tools from DEL are not more used in the KR community. Turning things around, I also ask what influence issues from KR and other fields will have on the future of DEL and related frameworks.

ABSTRACT. Autonomous multiagent systems are typically specified by using expressive properties representing, among others, the knowledge of the agents and what they can bring about collectively or on their own in a system. This tutorial will survey some of the current techniques addressing the model checking of multi-agent systems against agent-based specifications. Specifically, the tutorial will cover syntax and semantics for the temporal-epistemic logic CTLK and ATLK, the epistemic variant of alternating time temporal logic. It will also cover ordered- binary decision diagrams and symbolic labeling, parameter synthesis and strategy synthesis in ATLK specifications, verification of systems with an unbounded number of agents, and MCMAS: a model checker for multiagent systems implementing these techniques, as well as the application of the techniques to a range of scenarios including in service- oriented computing.

ABSTRACT. Ontology-based data access (OBDA) has recently become a hot research topic in knowledge representation and data management. The basic idea underlying OBDA is to superimpose an ontology (conceptual layer) to a set of data sources (data layer), and use the ontology as a virtual schema for querying the data sources. The ontology and the data sources are connected through declarative mappings that provide the semantic relationship between the two layers. This tutorial will introduce ontology-based data access and will deal with the problem of query answering in OBDA, focusing in particular on the query rewriting approach to query answering. After a brief history of ontology-based data access, the tutorial will overview the most important results obtained in this area. Then, the main query rewriting techniques for OBDA will be presented and compared. The relationship between OBDA and other research topics will also be addressed. The main focus of the tutorial is on ontologies expressed using Description Logics: however, other ontology specification languages, like Datalog+/–:, will be considered.

ABSTRACT. Autonomous multiagent systems are typically specified by using expressive properties representing, among others, the knowledge of the agents and what they can bring about collectively or on their own in a system. This tutorial will survey some of the current techniques addressing the model checking of multi-agent systems against agent-based specifications. Specifically, the tutorial will cover syntax and semantics for the temporal-epistemic logic CTLK and ATLK, the epistemic variant of alternating time temporal logic. It will also cover ordered- binary decision diagrams and symbolic labeling, parameter synthesis and strategy synthesis in ATLK specifications, verification of systems with an unbounded number of agents, and MCMAS: a model checker for multiagent systems implementing these techniques, as well as the application of the techniques to a range of scenarios including in service- oriented computing.

ABSTRACT. Ontology-based data access (OBDA) has recently become a hot research topic in knowledge representation and data management. The basic idea underlying OBDA is to superimpose an ontology (conceptual layer) to a set of data sources (data layer), and use the ontology as a virtual schema for querying the data sources. The ontology and the data sources are connected through declarative mappings that provide the semantic relationship between the two layers. This tutorial will introduce ontology-based data access and will deal with the problem of query answering in OBDA, focusing in particular on the query rewriting approach to query answering. After a brief history of ontology-based data access, the tutorial will overview the most important results obtained in this area. Then, the main query rewriting techniques for OBDA will be presented and compared. The relationship between OBDA and other research topics will also be addressed. The main focus of the tutorial is on ontologies expressed using Description Logics: however, other ontology specification languages, like Datalog+/–:, will be considered.