Tutorial 3. Tuesday, June 11th 14:00-17:30 (#102, 1F)

Non-Monotonic Reasoning on the Web

by Matteo Cristani (University of Verona, Italy)

Short Bios :

Matteo Cristani is Senior Researcher in the Department of Computer Science of the University of Verona. His expertis is the semantic web, and in particular, web applications with natural language processing and nonmonotonic reasoning. He has got numerous journal and conference papers in the mentioned communities. He has been the co-organised of the KDWeb (Knowledge Discovery on the WEB) workshop since 2017. The workshop has been a ICWE workshop in 2018.

Brief description : The development of the communities of Semantic Web and Web Engineering have been interlacing frequently in the past twenty years. Many interesting approaches to practical issues of web engineering involving reasoning techniques have been succesful, as happened, for instance, in the linked data recent investigations. A specific attention has also been posed on the basic issues of porting the web onto the semantic era as related to the development of XML-embedded languages, in particular OWL, as recently approached in many different applicative contexts.

Another community of research has deeply investigated the domain of web as related to reasoning, in a much more general way, and it is generally recognized as the reasoning web community. In this specific case, many studies have been dealing with problems related to change of knowledge, trustability and preference. In many senses, a unifying semantics of the borders among the three communities (OWL-related studies, semantic web engineering, reasoning with web data) can be given by rule-based reasoning systems also known as nonmonotonic reasoning.

Although the basic knwoledge on the above mentioned issues is common in practitioners and researchers of the various areas of web engineering, it is also true that a significant number of scholars and professionals, especially in the early stage of their careers can have a valuable advantage from a neater and wider survey on the nonmonotonic methods as applied to the web. The tutorial will be articulated in three parts:
  1. A general introduction to nonmonotonic reasoning systems, with special care for the defeasible reasoning framework. This is introduced in a formal way, by specifying syntax of rules, priorities and defeaters, and the derivation methods based on proof tags, associated with specific derivation and conclusion inferential methods, with the corresponding associated equivalent of proof conditions.
  2. A survey on the applications of the above to four typical issues of web reasoning:
(a) Solving conflicts in definitions given within reliable sources of knowledge. The example reported is Wikipedia, we show how to determine and possibly solve existing conflicts in definitions;

(b) Constructing debunking methods based on external reliable knowledge. The example reported is The skeptical inquirer, a scientific-oriented dissemination periodical;

(c) Providing decisions about trust on a given individual based on external sources. We show how the theoretical framework named Public announcement logic works;

(d) Ontology alignement with external sources of knowledge. The discovery of conflicts and their resolution using Rogets thesaurus.

Rule-based reasoning on the semantic web. An introduction to translation techniques and the solution of opinions, generalization, typicalization and other mitigative expressions into lambda-expressions for the translation of natural language onto rule-based semantic systems.