1.1 Obituary for Dr Harold Boley

Harold Boley studied in Hamburg and then worked at DFKI, the German Institute for Artificial Intelligence, from there he transitioned to the University of New Brunswick, where he taught Semantic Technologies, Declarative Programming, and Multi-Agent Systems and supervised graduate and undergraduate students. He also held a position at the National Research Council Canada. Harold was most well known for his work in the Web rule standards world (Boley et al. 2001). He chaired the RuleML Organization, maintained the RuleML Blog, and co-organized the annual RuleML Symposium, which was held without interruption since 2002, first as a workshop, and since 2005 as a stand-alone conference. In 2017, RuleML merged with Web Reasoning and Rule Systems conference into RuleML+RR. Besides his high involvement in the rule community, Harold contributed to various standardization efforts related to RuleML. Among the various RuleML achievements, he also contributed to OASIS LegalRuleML (Palmirani et al. 2011), as well as the publication of the W3C Rule Interchange Format (RIF) and the Semantic Web Rule Language (SWRL) (Horrocks et al. Reference Horrocks, Patel-Schneider, Boley, Tabet, Grosof and Dean2004). Harold’s research was in the area of Semantic Web rule languages and applications, with a particular focus on rule-based knowledge representation and reasoning. He was also involved in efforts joining rules and ontologies, for example, with sorted Horn logics in Relfun, with a hybrid combination of Datalog and Description Logics, and the Positional-Slotted Object-Applicative (PSOA) logic (Boley 2015), which integrates object-relational data and knowledge for rules and combines those with taxonomies (subclass ontologies). His most recent work focused on the translation of (PSOA) RuleML and Prolog with the PSOATransRun translator implementation and the visual rule language Grailog for visualizing PSOA RuleML (Thom et al. 2020).

1.2 RuleML+RR 2020 in numbers

This special issue presents selected papers from RuleML+RR 2020, which were further revised and extended. The technical program of the main track of RuleML+RR 2020 included 13 regular presentations of peer-reviewed research papers, which were selected from 30 submissions by a program committee consisting of nearly 60 academics. Additionally, the conference program included four invited talks, one of which was also represented by a full invited paper in the proceedings.

1.3 Overview of articles in this special issue

From the papers accepted to RuleML+RR 2020, a selection of four contributions with the highest reviewing scores and strong support was identified, and their authors were invited to submit an extended version of their papers to the Journal of Theory and Practice of Logic Programming. These articles underwent several rounds of a review process, during which each submission was reviewed by three referees; the authors incorporated the provided feedback, and the submissions were reviewed again until the reviewers were satisfied with the quality of the content.

“Swift Markov Logic for Probabilistic Knowledge Graphs” by Luigi Bellomarini, Eleonora Laurenza, Emanuel Sallinger and Evgeny Sherkhonov provides a framework for probabilistic reasoning in Vadalog-based Knowledge Graphs. The article starts with a general introduction, which describes the needed background and bridges the gap by explaining and giving the practical consequences of warded semantics. Accessibility of the material present in the article is also improved by the use of practical, intuitively understandable, examples from the banking domain. The approach has been tested on a number of real-world and synthetic KGs.

“Tackling the DM Challenges with cDMN: A Tight Integration of DMN and Constraint Reasoning” by Simon Vandevelde, Bram Aerts, and Joost Vennekens describes an extension to the Decision Model and Notation (DMN) standard, namely Constraint Decision Model and Notation (cDMN), with key features being constraint modeling, quantification, and the use of concepts such as types and functions. cDMN extends DMN with the purpose of allowing more complex domain knowledge to be represented, while retaining the usability of DMN. The proposed extension has been tested using complex problems and against other state-of-the-art DMN-like solvers, posted on the DM Community website. It has been shown that cDMN increases the expressiveness of DMN, allows solving more complex problems, and maintains a good level of readability and compactness.

“Declarative Approaches to Counterfactual Explanations for Classification” by Leopoldo Bertossi deals with the domain of Answer Set Programming, specifically it proposes answer-set programs that specify and compute counterfactual interventions as a basis for causality-based explanations to the outcomes from classification models, such as decision trees. The main contribution is the specification and computation of maximum-responsibility counterfactual explanations. The approach presented in the article also considers semantic and domain knowledge as well as probabilistic methods.

“Answering Fuzzy Queries over Fuzzy DL-Lite Ontologies” by Gabriella Pasi and Rafael Pealoza addressed the under-studied problem of answering conjunctive queries and threshold queries w.r.t. ontologies in fuzzy DL-Lite. The main contribution is a method for answering degree queries which works even if the TBox is graded, thus allowing for fuzzy TBox axioms. Despite its narrow topical focus, the article could appear to a wider audience based on the inclusion of multiple illustrative examples as well as the potential for the results to extend to other Horn description logics.