=Paper= {{Paper |id=None |storemode=property |title=None |pdfUrl=https://ceur-ws.org/Vol-1014/talk2.pdf |volume=Vol-1014 }} ==None== https://ceur-ws.org/Vol-1014/talk2.pdf
                 Actions, Processes, and Ontologies

                                 Giuseppe De Giacomo

              Dipartimento di Ingegneria Informatica, Automatica e Gestionale
                                Sapienza Università di Roma
                             Via Ariosto 25 00185 Roma, Italy
                           degiacomo@dis.uniroma1.it

Abstract
We overview reasoning about actions and processes over ontologies (or knowledge
bases) expressed in Description Logics (DLs). This is a critical research challenge that
has resisted good, robust solutions for a long time. In particular, while well-developed
theories of actions and processes exist in AI, e.g., the ones based on the Situation Cal-
culus, these theories are not well-behaved when applied to DL ontologies, since the
latter impose very difficult state constraints. Here we restate this difficulty, by showing
that combining even the simplest DLs and the simplest theory of actions in Situation
Calculus leads to undecidability of the simplest reasoning task: satisfiability. We then
look into a radically different approach, based on so-called Levesques functional view
of knowledge bases, that has been proved quite fruitful lately. This approach sees the
knowledge base (or ontology) as a system that allows for two kinds of operations: ASK,
which returns the (certain) answer to queries, and TELL, which produces a new knowl-
edge base as the result of the application of an atomic action (also related to update).
In particular we show that under this view even verification of sophisticated dynamic
properties (e.g., formulated in first-order variants of mu-calculus) over ontologies ex-
pressed in DLs, ranging from light-weight to very expressive ones, becomes decidable
under interesting general conditions.

Short CV. Giuseppe De Giacomo is a professor at Sapienza Università di Roma. He
has contributed to diverse areas of AI and CS. In the mid 90’s, he explored the cor-
respondence between Description Logics and Logics of Programs, and devised results
and reasoning techniques for expressive Description Logics, which contributed to the
birth of OWL and OWL2. At the end of the 90’, together with his group in Rome, he
started the research on conjunctive query answering in Description Logics. He is one of
the developers of the DL-Lite family of Description Logics, which have been recently
shaping the area of Ontology-Based Data Access. In Databases, he did foundational
work on data modeling, data integration, and view based query processing with regular
path queries, for both semi-structured and graph data. He has always had a deep interest
for Reasoning about Actions and Processes. He profoundly influenced the definition of
ConGolog and its successor IndiGolog, which are among the best known formalism
for expressing high-level robot programs. He devised one of the best known formal
approaches for service composition, the so-called “Roman Model”, which has been
further developed in AI for behavior compositions synthesis of devices and agents. Re-
cently, he provided foundational results on artifact-based business processes, showing
the decidability of verification and synthesis in infinite state data-aware systems.