The Semantics of a Description Logic Event Processing Query Language
Author(s): Riccardo Tommasini, Pieter Bonte, Emanuele Della Valle, Femke Ongenae, Filip De Turck
Full text: submitted version
Abstract: Stream Reasoning (SR) envisioned, investigated and proved the possibility to make sense of heterogeneous streaming data. SR requires techniques for querying and reasoning over time and, recently, approaches for reasoning about time. In our previous work, we proposed a more expressive stream reasoning framework called Description Logics Event Processing (DELP). DELP combines Complex Event Processing (CEP) and Description Logics (DL) to perform reasoning over time and about time. In this work, we investigate the foundational aspects of DELP. We define the query model, we provide a formal semantics that combines DL and a suitable temporal logic, i.e., DatalogMTL, and we explain DELP evaluation function.
Keywords: Complex Event Processing; Description Logics; Expressive stream reasoning; Temporal Logic
Review 1 (by anonymous reviewer)
(RELEVANCE TO ESWC) - (NOVELTY OF THE PROPOSED SOLUTION) - (CORRECTNESS AND COMPLETENESS OF THE PROPOSED SOLUTION) - (EVALUATION OF THE STATE-OF-THE-ART) - (DEMONSTRATION AND DISCUSSION OF THE PROPERTIES OF THE PROPOSED APPROACH) - (REPRODUCIBILITY AND GENERALITY OF THE EXPERIMENTAL STUDY) - (OVERALL SCORE) The paper proposes a formal semantics and query model for the Description Logic Event Processing framework using DatalogMTL. All in all, the paper makes the impression as if it was written in a haste. There are plenty of errors on the language level (to the degree that understanding is severely obstructed) as well as on the technical level. Also the general presentation lacks formality. Many definitions are vague and important ingredients (such as the semantics of DatalogMTL) are simply left away. I appreciate examples to explain central notions, but they should not replace the formal treatment. Most notably an equivalence proof for the (again very informally presented) translation from DELP to DatalogMTL_nr is totally missing.
Review 2 (by Zhisheng Huang)
(RELEVANCE TO ESWC) The paper addresses the event processing, one of the interesting issues in the Semantic Web. (NOVELTY OF THE PROPOSED SOLUTION) The paper proposes DELP, a framework for expressive streaming reasoning which combines Description Logics and Complex Event Processing. (CORRECTNESS AND COMPLETENESS OF THE PROPOSED SOLUTION) The paper contains some errors. For example, Line 7 in Listing 1.1 (Section 2.1 Knowledge Representation) is not correct. (EVALUATION OF THE STATE-OF-THE-ART) Related work have been discussed. (DEMONSTRATION AND DISCUSSION OF THE PROPERTIES OF THE PROPOSED APPROACH) The paper does not report any experiment with realistic data for an evaluation of the proposed approach. (REPRODUCIBILITY AND GENERALITY OF THE EXPERIMENTAL STUDY) It is unclear whether or not the proposed approach can deal with large scale data. (OVERALL SCORE) The paper addresses the event processing, one of the interesting issues in the Semantic Web. The paper proposes DELP, a framework for expressive streaming reasoning which combines Description Logics and Complex Event Processing. The paper does not report any experiment with realistic data for an evaluation of the proposed approach. It is unclear whether or not the proposed approach can deal with large scale data.
Review 3 (by Jens Lehmann)
(RELEVANCE TO ESWC) Stream Reasoning is an important topic of Semantic Web data processing. (NOVELTY OF THE PROPOSED SOLUTION) This technique is implemented using existing approaches as well as extending the authors previous works. The transformation of their query language into DatalogMTL provides a foundation and is supposed to be more approachable. The definitions could be expressed using RDF technology. (CORRECTNESS AND COMPLETENESS OF THE PROPOSED SOLUTION) Good effort has been made especially on the theoretical parts to define the DELP query language. Running examples makes the text easy to follow. (EVALUATION OF THE STATE-OF-THE-ART) A nice comparison table is presented, it could be further refined by discussing which other systems are based on DatalogMTL. Also some kind of remarks about integration would be welcome, for example can DELP make use of CSPARQL? Furthermore a comparison of different Complex Event definition systems could help to make a stronger point in favour of the proposed language. Finally, a reference to any ongoing standardisations would be helpful. (DEMONSTRATION AND DISCUSSION OF THE PROPERTIES OF THE PROPOSED APPROACH) The properties of the proposed approach are discussed and the two major reasons (readability, complexity in AC0) are given. (REPRODUCIBILITY AND GENERALITY OF THE EXPERIMENTAL STUDY) There was no experimental study. (and I cannot remove the dot) (OVERALL SCORE) Summary of the Paper ** The paper presented DELP, a query language for stream reasoning and complex event definition. Short description of the problem tackled in the paper, main contributions, and results** Authors define a query model, the formal semantics of their DELP temporal operators and the DELP query evaluation function. Strong Points (SPs) ** - comprehensive motivation and overview - easy examples - important topic for semantic web Weak Points (WPs) ** - no experiments or implementation are discussed - chapter Query Evaluation remains very high-level - why EPA++ is a good choice, could be explained Questions to the Authors (QAs) ** - Why ?room :equippedWith ?sensor on the entering/exit events? It is not clear why these events need any BGP - also see the questions given in Evaluation State-of-the-Art After rebuttal: I acknowledge the author response, but do not think the score needs to be adjusted.
Metareview by Jeff Z. Pan
The paper presented a framework for combing Description Logics and Complex Event Processing on top of an RDF Stream Processing Engine. The topic is a very important one for the Semantic Web. This is a contraversal paper -- the main concern is that the paper is lacking formal rigor. Indeed, reviewers strongly disagree with each other on the impact of this point on the overall assessment of this paper, even after discussions. With this view in mind, the submission was suggested to be a candidate cond-accept paper. Unfortunately, it didn't manage to stay above of the threshold in the final round of decision making. We hope the comments here are useful for a future submission of the work.