=Paper=
{{Paper
|id=Vol-201/paper-22
|storemode=property
|title=A Classification of Ontology Change
|pdfUrl=https://ceur-ws.org/Vol-201/03.pdf
|volume=Vol-201
|dblpUrl=https://dblp.org/rec/conf/swap/FlourisPA06
}}
==A Classification of Ontology Change==
https://ceur-ws.org/Vol-201/03.pdf
A Classification of Ontology Change
Giorgos Flouris1,2 , Dimitris Plexousakis2 , and Grigoris Antoniou2
1
Istituto della Scienza e delle Tecnologie della Informazione, C.N.R., Via G. Moruzzi, 1, 56124, Pisa, Italy
Email: georgios.flouris@isti.cnr.it
2
Institute of Computer Science, FO.R.T.H., P.O. Box 1385, GR 71110, Heraklion, Greece
Email: {dp, antoniou}@ics.forth.gr
Abstract— The problem of modifying an ontology in response causing unnecessary confusion as well as misunderstandings.
to a certain need for change is a complex and multifaceted one, The purpose of this paper is the introduction of a terminology
being addressed by several different, but closely related and often which follows the most common uses of the various terms
overlapping research disciplines. Unfortunately, the boundaries of
each such discipline are not clear, as certain terms are often used in the literature. Fixing this terminology will allow us to
with different meanings in the relevant literature. The purpose determine the boundaries of each field as well as to get a grip
of this paper is to identify the exact relationships, connections on their differences, overlaps, interactions and connections.
and overlaps between these research areas and determine the To do that, we perform a shallow, but broad, literature
boundaries of each field, by performing a broad review of the review on the field of ontology change, and introduce a broadly
relevant literature.
accepted terminology that will, hopefully, serve as a point of
I. I NTRODUCTION reference for the ontology change community. Our purpose is
Originally introduced by Aristotle, ontologies are often to give a clear overall picture of each relevant subfield and
viewed as the key means through which the Semantic Web determine the boundaries, interactions and overlaps between
vision [3] can be realized. Ontologies provide a means to the various areas; the interested reader is referred to the
formally define the basic terms and relations that comprise numerous bibliographic references that will appear throughout
the vocabulary of a certain domain of interest [34], enabling this paper for more details on each area or deeper results. A
machines to process information provided by human agents. comprehensive summary of the results of our survey can be
As a result, they can help in the representation of the content found in table I at the end of this paper.
of a web page in a formal manner so as to be suitable for use
II. O NTOLOGIES AND O NTOLOGY C HANGE
by an automated computer agent, search engine or other web
service. The importance of ontologies in current AI research A. What is an Ontology?
is also emphasized by the interest shown by both the research The term ontology has come to refer to a wide range of
and the enterprise community to various problems related to formal representations, including taxonomies, hierarchical ter-
ontologies and ontology manipulation [39]. minology vocabularies or detailed logical theories describing
Ontologies are often large structures, whose development a domain [44]. For this reason, a precise definition of the term
and maintenance give rise to interesting research problems. is rather difficult. A commonly used definition can be found
One of the most important such problems is the problem of in [21] where an ontology was defined to be a specification of
modifying an ontology in response to a certain need. In this a shared conceptualization of a domain.
paper, the term ontology change will be used to describe this A more formal, algebraic, approach, identifies an ontology
problem; the term will be used in a broad sense, covering any as a pair < S, A >, where S is the signature of the ontology
type of change, including changes to the ontology in response (being modeled by some mathematical structure, such as a
to external events, changes dictated by the ontology engineer, lattice, a poset or an unstructured set) and A is the set of
changes forced by heterogeneity considerations and so on. ontological axioms, which specify the intended interpretation
In order to cope with the complex problem of ontology of the signature in a given domain of discourse [27].
change, several related research disciplines have emerged
(such as ontology evolution, alignment, merging, mapping B. Ontology Change
etc), each dealing with a different facet of the problem. These Several reasons for changing an ontology have been iden-
areas are greatly interlinked; as a result, several works and tified in the literature. An ontology, just like any structure
systems deal with more than one of these topics causing a storing information, may need to change simply because the
certain confusion to a newcomer. This confusion is further modeled domain has changed [55]; but even if we assume a
increased by the fact that certain terms are often used with static domain, which is a rather unrealistic assumption for most
different meanings in the relevant literature, denoting similar, applications, we may need to change the perspective under
but not identical, research directions or concepts. For examples which the domain is viewed [44], or we may discover a design
of such confusing and overused terms refer to [13], [51]. flaw in our original conceptualization [52]; we may also wish
We believe that this lack of a standard terminology consti- to adapt to a change in users’ needs or perspective and/or
tutes a major bottleneck for the ontology change community, incorporate additional functionality [22]; new information,
previously unknown, classified or otherwise unavailable may viewpoint of the conceptualization, information received by
become available or different features of the domain may some external source, a change in the domain, communica-
become important [25]. tion needs between heterogeneous sources of information or
In addition, ontology development is becoming more and ontologies, the fusion of information from different ontologies
more a collaborative and parallelized process, whose subprod- and so on.
ucts need to be combined to produce the final ontology [32]; This definition covers several related research areas which
this process would require changes in each subontology to are studied separately in the literature. In this paper, we
reach a consistent final state. But even then, the so-called identify nine such areas, namely ontology mapping, morphism,
“final” state is rarely final, as ontology development is usually alignment, articulation, translation, evolution, versioning, in-
an ongoing process [25]. tegration and merging. Each of these areas deals with a certain
The complex web of dependencies that is usually formed facet of the problem from a different view or perspective,
around an ontology is another common reason for change. covering different application needs, change scenarios or needs
The distributed nature of the Semantic Web implies that the for change (see table I for a comprehensive summary).
knowledge engineer has no control over dependent and/or These fields are greatly interlinked, so several papers deal
depending ontologies; if any of these ontologies change, the with more than one of these problems. In other cases, the same
local ontology might also need to be modified [25]. In other term is used in different papers to describe different research
cases, a certain agent, service or application may need to use areas. This situation can easily lead to misunderstandings,
an ontology whose terminology or representation is different confusion and unnecessary waste of effort, especially for
from the one it can understand [9], so he needs to perform a newcomer. In the following sections, we will attempt to
some kind of translation (change) in the imported ontology. precisely define the boundaries of each area and uncover their
Finally, we may need to merge or integrate information from relations, overlaps and differences. This attempt will hopefully
two or more ontologies in order to produce a more appropriate draw a fine line between the various research areas, allowing
one for some application [51]. the clarification of the meaning of each term and making
Several philosophical problems related to knowledge update the differences and similarities between them explicit. The
in general have been identified in the research area of belief definitions provided here will not be arbitrary, but will be
revision [19], [20], [28]; many of them are also applicable to based on the most common uses of each term in the literature.
knowledge represented in ontologies [12], [13]. However, the
III. O NTOLOGY E VOLUTION AND V ERSIONING
problem is further complicated by the large size of modern
day ontologies [39] and by the aforementioned ontology A. Disambiguating the Terms
interdependencies; even subtle changes in an ontology may Ontology versioning is often considered a stronger variant
have unforeseeable effects in dependent and/or depending of ontology evolution [23]. Under that viewpoint, ontology
applications, services, data and other ontologies [54]. evolution is the process of changing an ontology without losing
These facts raise the need to maintain different interoperable data or negating its validity, whereas ontology versioning
versions of the same ontology [25], [26], [31], a problem should additionally guarantee the validity, interoperability and
greatly interwoven with ontology change [30]. Moreover, management of all previous versions, including the current
heterogeneity leads to problems when an agent, service or one, as well as transparent access to these versions.
application uses information from two different ontologies [9]. This viewpoint is influenced by related research on re-
As ontologies often cover overlapping domains using different lational and object-oriented database schema evolution and
viewpoints and terminology, some kind of translation may be versioning [18], [29], [50]. A survey on the differences and
necessary in many practical applications. similarities of ontologies and databases, as well as their impact
All these arguments indicate the importance of the problem with respect to evolution and versioning, can be found in
of ontology change and motivate us to use the term in order [44]. In this paper it is argued that ontology evolution and
to cover all aspects of ontology modification, as well as the versioning become indistinguishable under this understanding,
problems that are indirectly related to the change operation because, due to the distributed nature of the Semantic Web,
such as the maintenance of different versions of an ontology multiple versions of ontologies are bound to exist and must
or the translation of ontological information in a common be supported. Furthermore, ontologies and dependent elements
terminology. More specifically, we will use the term ontology are likely to be owned by different parties; as a result, some
change to refer to the problem of deciding the modifications parties may be unprepared to change and others may even
to perform upon an ontology in response to a certain need be opposed to it [25]. All these facts force us to maintain
for change as well as the implementation of these modifica- and support different versions of ontologies, making ontology
tions and the management of their effects in depending data, evolution (under this understanding) useless in practice.
ontologies, services, applications, agents or other elements. We believe that the problem of modifying the ontology (on-
In this definition, the need to change the ontology may tology evolution) should be clearly separated from the problem
take several different forms, including, but not limited to, the of maintaining the interoperability of different versions of the
discovery of new information (some new instance data, another ontology (ontology versioning). This distinction is not always
ontology, a new observation etc), a change in the focus or the clear in the literature, because the ontology dependencies and
interrelationships force us to consider the issue of propagating replaced by a series of atomic changes. Even though possible,
the changes to dependent elements [37]. This tight coupling it is not generally appropriate to use a series of atomic changes
has caused ontology evolution algorithms to deal with these to replace a complex change, as this might cause undesirable
problems as well. For example, in [54], ontology evolution is side-effects [54]; the proper level of granularity should be
defined as the timely adaptation of an ontology to changed identified at each case. Unfortunately, there is no general
business requirements, to trends in ontology instances and consensus in the literature on the type and number of complex
patterns of usage of the ontology-based application, as well as changes that are necessary. In [54], 12 different complex
the consistent management and propagation of these changes changes are identified; in [44], 22 such operations are listed; in
to dependent elements. [56] however, the authors mention that they have identified 120
On the contrary, here we define ontology evolution to refer different interesting complex operations and that the list is still
to the process of modifying an ontology in response to a growing! In fact, the number of definable complex operations
certain change in the domain or its conceptualization [13]; can only be limited by setting a granularity threshold on the
on the other hand, ontology versioning refers to the ability operations considered; if we allow unlimited granularity, we
to handle an evolving ontology by creating and managing will be able to define more and more operations of coarser
different versions of it [30]. Thus, ontology evolution is and coarser granularity, limited only by our imagination [32].
restricted to the process of modifying an ontology while Thus, creating a complete list of complex operations is not
maintaining its validity, whereas ontology versioning deals possible, but, fortunately, it is not necessary either, since a
with the problem of managing different versions of an evolving complex operation can always be defined as a series of atomic
ontology, maintaining interoperability between versions and operations [32].
providing transparent access to each version as required by The third phase is the semantics of change phase, in which
the accessing element. we identify and address any problems that will be caused when
the required changes are actually implemented, thus guarantee-
B. Ontology Evolution: General Discussion ing the validity of the ontology at the end of the process. For
Since an ontology is a specification of a shared conceptual- example, if a concept is deleted, we need (among other things)
ization of a domain [21], a change may be caused by either a to determine what to do with its instances (e.g., delete them or
change in the domain, a change in the conceptualization or a re-classify them). In [54], it is suggested that the final decision
change in the specification [30]. Changes in the specification should be made indirectly by the ontology engineer, through
refer to changes in the way the conceptualization is formally the selection of certain pre-determined evolution strategies,
recorded, i.e., changes in the representation language. This indicating the appropriate action in each case. Other (manual
type of change is dealt with in the field of ontology translation or semi-automatic) approaches are also possible (see [23]).
(see the next section and table I at the end of this paper). Thus, This phase is probably the most crucial of ontology evolution,
our definition of ontology evolution covers the first two types because during that phase the direct and indirect effects of a
of change only (domain and conceptualization changes). given change request are determined.
Both types of changes are not rare. The conceptualization The change implementation phase follows, where the
of the domain may change because of a new observation changes are physically applied to the ontology, using an
or measurement, a change in the viewpoint or usage of appropriate tool, like, for example, the KAON API [54]. Such
the ontology, newly-gained access to information that was a tool should have transactional properties, based on the ACID
previously unknown, classified or otherwise unavailable and so model, i.e., guaranteeing Atomicity, Consistency, Isolation and
on. The domain itself may also change, as the real world itself Durability of changes [23]. It should also present the changes
is generally not static but evolves over time. More examples to the ontology engineer for final verification and keep a log
of reasons initiating changes can be found in [30], [44]. of the implemented changes [23].
The implemented changes need to be propagated to all
C. Ontology Evolution Phases interested parties; this is the role of the change propagation
In order to tame the complexity of the problem, six phases phase. In [37], two different methods to address the problem
of ontology evolution have been identified, occurring in a are compared, namely push-based and pull-based approaches.
cyclic loop [54]. Initially, we have the change capturing phase, Under a push-based approach, the changes are propagated
where the changes to be performed are identified. Three types to the dependent ontologies as they happen; in a pull-based
of change capturing have been distinguished: structure-driven, approach, the propagation is initiated only after the explicit
usage-driven and data-driven [23]. request of each of the dependent elements. In both [37]
Once the changes have been determined, they have to and [54] the push-based approach is favored. Alternatively,
be properly (and formally) represented during the change one could avoid this step altogether, by using an ontology
representation phase. There are two major types of changes, versioning algorithm [31], allowing the interested parties to
namely atomic and complex [56] (also called elementary and work with the original version of the ontology and update to
composite in [54]). Atomic changes represent simple, fine- the newer version at their own pace, if at all. This alternative
grained changes such as the deletion of a concept. Complex is considered more realistic for practical purposes [25].
changes represent more coarse-grained changes and can be Finally, the change validation phase allows the ontology
engineer to review the changes and possibly undo them. This based on some type of identification mechanism to differen-
phase may uncover further problems with the ontology, thus tiate between various versions of an ontology, but it is not
initiating new changes that need to be performed to improve always clear when two ontologies constitute different versions.
the conceptualization; in this case, we need to start over Should any change in the file that stores the ontology constitute
by applying the change capturing phase of a new evolution the creation of a new version? When a concept specification
process, closing the cyclic loop. changes, but the new specification is semantically equivalent
Notice that heterogeneity issues are not handled by the to the original one, does this constitute a new version? More
above ontology evolution model. Obviously, any approach generally, when the ontology changes syntactically, but not
to ontology evolution would collapse in the presence of semantically, does this constitute a new version? These and
heterogeneity, unless coupled with some algorithm that deals similar problems are dealt with in [25], [31].
with heterogeneity (like the ones discussed in the next section). Another desirable property of an ontology versioning sys-
However, under the proposed model, this is not a problem, as tem is the ability to allow transparent access to different
the ontology engineer identifies the changes to be performed versions of the ontology, by automatically relating versions
during the change representation phase, so it can be reasonably with dependent elements [30]. Other issues involved is the so-
assumed that these changes will be represented in a suitable called “packaging of changes” [31] as well as the different
terminology. An alternative model of ontology evolution, in- types of compatibility and how these are identified [30].
volving five phases, has been proposed in [52]. Another related problem is the introduction of a certain
version relation between ontological elements (such as classes)
D. The Current State of the Art in Ontology Evolution that appear in different versions of the ontology and the
properties that such a relation should have. This relation is
The current state of the art in ontology evolution, as well
called a change specification in [30] and its role is to make the
as a list of relevant tools can be found in [23]. Some of
relationship between different versions of ontological elements
these tools are simple ontology editors, whereas others provide
explicit. Using this relation, one can identify the changes that
more specialized features to the user. In some cases, the user
any given element went through between different versions; in
can define some kind of pre-defined evolution strategies [54]
addition, a version relation should include certain meta-data
that control how changes will be made, thus allowing the
regarding these changes [31]. In [52] this relation is stored
tool to perform some of the required changes automatically.
using a version log which is actually a specially designed
Other tools allow collaborative edits, i.e., several users can
ontology containing the different versions of each element,
work simultaneously on the same ontology [7], whereas others
as well as the relation between them and some related meta-
support transactional changes [23]. In other works, features
data. Similar considerations led to the definition of migration
related to ontology versioning, undo/redo operations and other
specifications [60], which associate concepts between different
helpful utilities are supported [7]. Some tools provide intuitive
versions of an ontology after a change has been performed.
graphical interfaces that help the visualization of the process
[33]. For more details on such systems refer to [7], [23]. F. The Current State of the Art in Ontology Versioning
A declarative language for changing the data portion of an
As an aid to the task of ontology versioning, certain
RDF ontology appears in [38]. An alternative approach that
tools have been developed which automatically identify the
uses belief revision techniques to handle ontology evolution
differences between ontology versions; unfortunately, most
has recently appeared [11], [14]–[16]; similar approaches, at a
such tools provide information at the level of atomic changes
preliminary stage, appear in [35], [40]. An interesting variation
[32]. PROMPTDIFF [45] uses certain heuristics to compare
of the problem appears in [17], [58], [59], where the evolving
different versions of ontologies and outline their differences,
objects (and therefore the main objects of study) are the
by producing a structural diff between them. OntoView [31]
concepts; this viewpoint is quite different from the standard
contains a tool similar to PROMPTDIFF, whose output is a
one, in which the evolving object is an ontology as a whole.
certain ontology of changes.
A survey on the different ways that can be used to rep-
E. Ontology Versioning
resent a set of changes, as well as the relation and possible
Once the actual changes have been performed, ontology interactions between such representations can be found in [32];
versioning comes into play. Ontology versioning typically in the same paper, another ontology of changes is proposed,
involves the storage of both the old and the new version of containing both atomic and complex operations. A similar
the ontology and takes into account identification issues (i.e., ontology of changes is proposed in [52], where the changes
how to identify the different versions of the ontology), the are identified through a version log stored in this ontology of
relation between different versions (i.e., a tree of versions changes.
resulting from the various ontology modifications) as well A method to identify compatibility between versions is
as compatibility information (i.e., information regarding the presented in [24], [25] where the SHOE language [36] is used
compatibility of any pair of ontology versions). to make backward compatibility between versions explicit and
Several non-trivial problems are associated with this task. determinable by a computer agent. This is an indirect approach
For example, any ontology versioning algorithm should be to the problem of ontology versioning, because it allows the
computer agent to determine autonomously which version that relate both ontological signatures and axioms. Notice that
to use, as opposed to [30], [31], where a more direct and ontology morphism, unlike the other fields discussed in this
centralized path is taken. In [26], a temporal logic approach section, is not restricted to the ontology signature only, but
is used to allow access in different versions of an ontology. covers the ontological axioms as well.
In ontology mapping and morphism the ontologies are
IV. O NTOLOGY M APPING , M ORPHISM , A LIGNMENT,
related via functions; an interesting, and more general, alterna-
A RTICULATION AND T RANSLATION
tive is by means of a relation. The task of finding relationships
A. General Discussion between signature entities belonging to two different ontolo-
Work related to these areas tries to mitigate the problems gies is called ontology alignment. So the output of ontology
caused by the heterogeneity of the Semantic Web. The general alignment is a binary relationship between the ontological
motivation for these research fields is that different ontologies signatures. This approach is more liberal, allowing greater
(and sources of information based upon different ontologies) flexibility, so it is more commonly used in practice.
generally use different terminology, different representation A binary relationship could be decomposed into a pair of
languages and different syntax to refer to the same or similar total functions from a common intermediate source; therefore,
concepts. A nice list of use cases where this heterogeneity the alignment of two ontologies could be described by means
may cause problems can be found in [9]. of a pair of ontology mappings from a common intermediate
The obvious solution to this problem is the provision of ontology. We use the term ontology articulation to refer to the
a set of translation rules of some kind that will allow us process of determining the intermediate ontology and the two
to nullify these terminological differences. To put it simply, mappings to the initial ontologies.
the goal of the whole process is to make two ontologies Finally, the term ontology translation is used in the literature
refer to same entities using the same name and to differ- with two different meanings. Under one understanding, ontol-
ent entities using different names. For example, we should ogy translation refers to the process of changing the formal
be able to identify that the concepts RESEARCHER and representation of the ontology from one language to another.
RESEARCH STAFF MEMBER that appear in two different This changes the syntactic (only) form of the axioms, but not
ontologies refer to the same real-world concept, i.e., the class the signature of the ontology. Under the second understanding,
of researchers. We should also be able to differentiate between ontology translation refers to a translation of the signature, in
two different uses of the entity CHAIR, as it could refer to a manner similar to that of ontology mapping. The difference
the class of chairs (as a furniture) in one ontology and to the between ontology mapping and ontology translation is that the
people forming a Workshop’s Chair in another. former specifies the functions that relate the two ontologies’
Even though these research fields basically deal with the signatures, whereas the latter applies these functions to actu-
same problem (i.e., heterogeneity resolution), they can be ally implement the mapping.
identified based on the type of translation rules that is produced
C. Methodology and the Current State of the Art
at the output. Due to the close relationship between these
areas, sometimes the term ontology alignment (e.g., in [9]) or The methods commonly used to address the problem of
ontology mapping (e.g., in [27]) is used to refer collectively heterogeneity include studying the taxonomic or mereological
to all of them. In this section, we will try to disambiguate the structure of the entities, evaluating name similarities (where
situation; most of the material for this section is taken from the names are compared as strings) and so on. Other methods
[9] and [27]. use a thesaurus to study the linguistic similarities of names,
use semantic approaches, or determine the similarity based
B. Definitions on the instances of each entity. The final similarity evaluation
The term ontology mapping refers to the task of relating may also be affected by the evaluation of the similarity of the
the signatures of two ontologies that share the same domain entities’ neighborhood. In real systems, a combination of some
of discourse in such a way that the mathematical structure of these approaches with some kind of human intervention
of ontological signatures and their intended interpretations, usually works best. A detailed classification and description
as specified by the ontological axioms, are respected. The of these methods can be found in [9].
result of an ontology mapping algorithm is a collection of Two popular systems that deal with heterogeneity are
functions on ontological signatures. A similar (and equivalent) PROMPT [46], [47] (originally called SMART [48]) and
definition appears in [4], where ontology mapping is defined as Chimaera [39]. In [10], the term ontology matching is used to
a (declarative) specification of the semantic overlap between refer to an ontology mapping algorithm based on the linguistic
two ontologies, which can be either one-way (injective) or properties of terms, using a thesaurus based on WordNet [41].
two-way (bijective). In [53], a certain string metric is proposed to evaluate name
This definition restricts the mappings to ontological signa- similarities of elements in different ontologies, upon which an
tures. A more ambitious and interesting approach would be to ontology alignment algorithm could be based. Some thoughts
create mappings that deal with both the signatures and the ax- on the issue of heterogeneity in the context of the SHOE
ioms of the ontologies. The term ontology morphism refers to language can be found in [24], [25]. An interesting method of
that approach, i.e., the development of a collection of functions improving the results of an alignment process, which exploits
user validation combined with machine learning techniques, In [46], [47] ontology merging is defined as the process
can be found in [8]. of creating a new, coherent ontology that includes information
In [42], a probabilistic technique is used towards this from two or more source ontologies; this is implicitly assumed
aim; the final similarity evaluation of this ontology mapping to include the process of resolving any possible heterogeneities
algorithm is affected by the similarity probabilities of each between the merged ontologies. In these papers, ontology
entity’s neighborhood, improving the initial mapping result. merging and alignment are understood as variations of the
Another method based on probabilistic analysis, which takes same problem, the only difference being that ontology merging
into account uncertainty issues in the mapped ontologies can results in the creation of a new ontology, whereas in ontology
be found in [49]. A general-purpose approach to the problem alignment the merged ontologies persist, with links established
of translation is described in [6]. A much more extensive list between them.
of systems and works related to these research areas can be A similar use of the term can be found in [39], whereas,
found in [4], [9], [27]; a relevant evaluation appears in [1]. in [25], the same research area is described using the term
Unfortunately, heterogeneity resolution in ontologies still ontology integration. According to [35], ontology merging
relies on human intervention; however, the process has to amounts to making sure that different agents use the same
be automatic in order to be practical [27]. In this direction, terms in identical ways (in a manner similar to ontology
advances in the field of natural language processing will alignment). In [27] ontology integration is defined as the
probably help researchers gain a better understanding on the process of combining ontologies to build new ones, but whose
processes behind automatic heterogeneity resolution [27]. respective signatures are usually not interpreted in the same
domain of discourse. In [5] the same term is used to refer
D. Heterogeneity Resolution and Ontology Change
to the process of combining a number of local ontologies in
Notice that most of the fields studied in this section do order to build a global one, with the purpose of being able to
not directly modify any ontology, but provide translation answer queries over the local ontologies using the global one
rules that relate ontologies. As a result, many would argue and the mappings between these ontologies.
that these research areas should not be considered subfields Here, we will define these terms along the lines of [51],
of ontology change. We believe otherwise, for two reasons. which was an attempt to disambiguate between different uses
First, heterogeneity resolution constitutes a prerequisite for of the term ontology integration. Three different uses of
successful ontology change, as it makes no sense to try to the term were identified in that paper. The first refers to
change an ontology in response to new information unless both the composition of ontologies covering loosely related (i.e.,
the ontology and the new information are formulated using the similar) domains; this is mainly used when building a new
same terminology, language and syntax. So, it makes practical ontology that covers all these domains. The term ontology
sense to study these fields along with the problem of ontology integration has been reserved for this process.
change. The second use of the word refers to the combination of
Second, heterogeneity resolution implicitly requires the ontologies covering highly overlapping or identical domains;
modification of an ontology, so it is really a subfield of this process is used to fuse ontologies that contain information
ontology change in the wide sense of the term used in about the same subject into one large (and hopefully more
this paper. Indeed, consider two agents with heterogeneous accurate) ontology. The term ontology merging was attached
ontologies that need to communicate and some translation to this interpretation.
rules allowing this communication. In this particular example, Finally, the third use of the term integration refers to the
the “need for change” is the need for communication. The development of an application that uses one or more ontolo-
rules produced do not directly modify any ontology; however, gies; the more appropriate term ontology use was reserved for
they allow each agent to change the other agent’s ontology this process. In this paper, we focus on the first two research
locally to fit his own terminology, language and syntax. So areas, namely ontology integration and merging.
the change in this case is made on-the-fly by each agent. In
this sense, we could consider ontology mapping and the other B. Differences Between Integration and Merging
fields studied in this section to be subfields of ontology change There are certain subtle differences between the processes
that simply provide us with a method to change an ontology of integration and merging. Ontology integration is mainly
(even though no change is performed explicitly). applied when the main concern is the reuse of other ontologies.
The domain of discourse of the new ontology is usually more
V. O NTOLOGY I NTEGRATION AND M ERGING
general than the domain of any of the source ontologies and
A. Discussion and Definitions integration often places the different (source) ontologies in
Both ontology integration and merging refer to the construc- different modules that comprise the resulting ontology.
tion of a new ontology based on the information found in two On the other hand, in ontology merging, the focus is on
or more source ontologies; yet, the two terms refer to slightly creating an ontology that combines information on a given
different research areas. Unfortunately, the exact meaning of topic from different sources. In this case, the information from
each term is not clear in the literature, as they are often used the source ontologies is greatly intermingled, so it is difficult
interchangeably [51], causing a certain amount of confusion. to identify the part(s) of the final ontology that resulted from
each source ontology. A more detailed discussion can be found ontologies should be merged in the resulting ontology. The
in [51]. final choice relies on the ontology engineer. Some ideas on
ontology merging (called integration there) in the context of
C. Integration, Merging and Heterogeneity Resolution the SHOE language can be found in [25]; however, [25] is
It is a common practice in the literature (e.g., [4], [25], [46], focused on the part of merging that deals with heterogeneity
[51]), to consider heterogeneity resolution to be an internal resolution. In [5], an interesting theoretical framework for
part of ontology merging or integration. This is a reasonable ontology integration is defined, focusing on the creation of
choice, because in most cases the fused ontologies come mappings between the source and the resulting ontologies and
from different sources, so they are generally heterogeneous how these mappings can be exploited for query answering.
in terms of vocabulary, syntax, representation etc. Therefore, An interesting theoretical approach to ontology merging can
the task of resolving any heterogeneities between the source be found in [2], whereas in [47] some interesting connections
ontologies constitutes a major part of the task of ontology of object-oriented programming with the problem of ontology
merging (or integration). This is mostly true in merging, where merging are uncovered. The FCA-MERGE algorithm [57]
the domain of discourse is (almost) identical. This has led to performs ontology integration in a very efficient way, but is
even more confusion on the exact meaning of the terms, as based on certain strong assumptions. A more detailed list of
several researchers consider ontology merging (or integration) tools and systems related to the problem can be found in [4],
and alignment to be variations of the same problem (e.g., [35], [51].
[46]). Even though the problem of evaluating ontology merging
However, it should be clear that simply resolving the techniques is still open in AI [57], certain comparison attempts
heterogeneity issues between two ontologies is not sufficient have been made. In [34], the authors perform a comparison
for successful integration (or merging); recall that different between PROMPT and Chimaera in the context of bioinfor-
ontologies may encode different viewpoints regarding the real matics. In [46], the same two tools are compared with the
world, thus several conceptual differences are bound to exist, generic Protégé-2000 [43]. Furthermore, [39] compares the
even if the same terminology is used. This is reminiscent of efficiency of ontology merging with a simple plain-text editor,
how beliefs held by different people are often different (and merging with the Ontolingua editor and merging with the
in some cases contradictory), even if a common terminology specialized tool Chimaera, which is described in the same
is agreed upon. paper. These comparisons are made from a certain standpoint;
Similarly, modeling conventions and choices may be dif- a general, objective comparison is difficult, as it is not clear
ferent; one example of modeling choice that often depends how the utility of such tools could be measured [39].
on personal taste or convention is whether to model a certain
distinction between similar elements by introducing separate VI. C ONCLUSION
classes or by introducing a qualifying attribute relation in In this paper, we performed a shallow, but broad literature
one class [6]. Such modeling differences need to be taken review covering all the diverse types of ontology change. This
into account when selecting what to keep from each ontology allowed us to fix a terminology in an area that is plagued
during the integration or merging process. Reckless inclusion by underspecified and confusing terms which are used with
of ontology elements and axioms from the source ontologies different meanings by different researchers. This terminology
(even when homogeneous) is likely to lead to a problematic, was not introduced in an arbitrary manner, but was based on
invalid or inconsistent ontology. similar previous attempts (like [27], [51]) and on the most
common uses of the terms in the literature. We hope that
D. State of the Art in Ontology Integration and Merging
our work will prove helpful towards the clarification of the
According to [6], the process of merging can be broken boundaries and relations between the various fields and will
down in five steps. During the first step, we identify the serve as a starting point for researchers interested in any of
semantic overlap between the source ontologies; during the the many facets of ontology change. A summary of the results
second, we devise ways (transformations) to bring the sources of our study can be found in table I at the end of this paper.
into mutual agreement in terms of terminology, representation
etc. In the third step, we apply these transformations, so we can ACKNOWLEDGMENT
now take the union of the sources (fourth step). The final step The authors are grateful to Panos Constantopoulos, Vassilis
consists of evaluating the resulting ontology for consistency, Christophides and Nicolas Spyratos for helpful comments in
uniformity, redundancy, quality of conceptualization etc; this an earlier draft of this work. This work was carried out during
evaluation might force us to repeat some or all of the above the first author’s tenure of an ERCIM “Alain Bensoussan”
steps. The tool described in [6] facilitates the design and Fellowship Programme.
implementation of the transformations used in the merging
process (second and third step). R EFERENCES
The main tools used for ontology merging are PROMPT
[1] P. Avesani, F. Giunchiglia, M. Yatskevich. A Large Scale Taxonomy
[46] and Chimaera [39]. These tools use a semi-automatic Mapping Evaluation. In Proceedings of the 4th International Semantic
approach focused on suggesting how elements from the source Web Conference (ISWC-05), pp. 67-81, 2005.
[2] T. Bench-Capon, G. Malcolm. Formalizing Ontologies and Their Rela- [26] Z. Huang, H. Stuckenschmidt. Reasoning with Multi-version Ontologies:
tions. In Proceedings of the 16th International Conference on Database A Temporal Logic Approach. In Proceedings of the 4th International
and Expert Systems Applications (DEXA-99), pp. 250-259, 1999. Semantic Web Conference (ISWC-05), pp. 398-412, 2005.
[3] T. Berners-Lee, J. Hendler, O. Lassila. The Semantic Web. Scientific [27] Y. Kalfoglou, M. Schorlemmer. Ontology Mapping: the State of the Art.
American, 284(5):34-43, 2001. Knowledge Engineering Review, 18(1), pp. 1-31, 2003.
[4] J. de Bruijn, F. Martin-Recuerda, D. Manov, M. Ehrig. D4.2.1: State of [28] H. Katsuno, A.O. Mendelzon. On the Difference Between Updating
the Art Survey on Ontology Merging and Aligning. SEKT Deliverable, a Knowledge Base and Revising It. Technical Report on Knowledge
2004. Representation and Reasoning, Univ. of Toronto, KRR-TR-90-6, 1990.
[5] D. Calvanese, G. De Giacomo, M. Lenzerini. A Framework for Ontology [29] W. Kim, H.T. Chou. Versions of Schema for Object-Oriented Databases.
Integration. In I. Cruz, S. Decker, J. Euzenat, D. McGuinness (eds). The In Proceedings of the 14th International Conference on Very Large Data
Emerging Semantic Web. Selected Papers from the First Semantic Web Bases (VLDB-88), pp. 148-159, 1988.
Working Symposium, pp. 201-214. IOS Press, 2002. [30] M. Klein, D. Fensel. Ontology Versioning on the Semantic Web. In
[6] H. Chalupsky. OntoMorph: A Translation System for Symbolic Knowl- Proceedings of the International Semantic Web Working Symposium,
edge. In Proceedings of the 7th International Conference on Knowledge 2001.
Representation and Reasoning (KR-00), 2000. [31] M. Klein, D. Fensel, A. Kiryakov, D. Ognyanov. Ontology Version-
[7] A.J. Duineveld, R. Stoter, M.R. Weiden, B. Kenepa, V.R. Benjamins. ing and Change Detection on the Web. In Proceedings of the 13th
WonderTools? A Comparative Study of Ontological Engineering Tools. International Conference on Knowledge Engineering and Knowledge
International Journal of Human-Computer Studies, 52(6):1111-1133, Management, 2002.
2000. [32] M. Klein, N.F. Noy. A Component-Based Framework for Ontology
[8] M. Ehrig, S. Staab, Y. Sure. Bootstrapping Ontology Alignment Methods Evolution. In Proceedings of the IJCAI-03 Workshop on Ontologies and
with APFEL. In Proceedings of the 4th International Semantic Web Distributed Systems, CEUR-WS, vol. 71, 2003.
Conference, 2005. [33] S.C. Lam, D.H. Sleeman, W. Vasconselos. ReTAX++: A Tool for
[9] J. Euzenat, T. Le Bach, J. Barrasa, P. Bouquet, J. de Bo, R. Dieng, Browsing and Revising Ontologies. In Poster Proceedings of the 4th
M. Ehrig, M. Hauswirth, M. Jarrar, R. Lara, D. Maynard, A. Napoli, International Semantic Web Conference (ISWC-05), PID-33, 2005.
G. Stamou, H. Stuckeschmidt, P. Shvaiko, S. Tessaris, S. van Acker, [34] P. Lambrix, A. Edberg. Evaluation of Ontology Merging Tools in
I. Zaihrayeu. D2.2.3: State of the Art on Ontology Alignment. SEKT Bioinformatics. In Proceedings of the 8th Pacific Symposium on Bio-
Deliverable, 2004. computing, pp. 589-600, 2003.
[10] A. Ferrara. Methods and Techniques for Ontology Matching and Evo- [35] K. Lee, T. Meyer. A Classification of Ontology Modification. In Proceed-
lution in Open Distributed Systems. In Proceedings of the 16th Interna- ings of the 17th Australian Joint Conference on Artificial Intelligence,
tional Conference on Advanced Information Systems Engineering, 2004. 2004.
[11] G. Flouris. On Belief Change and Ontology Evolution. Doctoral Disser- [36] S. Luke, L. Spector, D. Rager, J. Hendler. Ontology-based Web Agents.
tation, Department of Computer Science, Univ. of Crete, 2006. In Proceedings of the 1st International Conference on Autonomous
[12] G. Flouris, D. Plexousakis. Bridging Ontology Evolution and Belief Agents, 1997.
Change. In Proceedings of the 4th Hellenic Conference on Artificial [37] A. Maedche, B. Motik, L. Stojanovic, R. Studer, R. Volz. An Infras-
Intelligence, 2006. tructure for Searching, Reusing and Evolving Distributed Ontologies.
[13] G. Flouris, D. Plexousakis. Handling Ontology Change: Survey and In Proceedings of the 12th International World Wide Web Conference,
Proposal for a Future Research Direction. Technical Report FORTH- 2003.
ICS/TR-362, September 2005. [38] M. Magiridou, S. Sahtouris, V. Christophides, M. Koubarakis. RUL:
[14] G. Flouris, D. Plexousakis, G. Antoniou. Evolving Ontology Evolution. A Declarative Update Language for RDF. In Proceedings of the 4th
In Proceedings of the 32nd International Conference on Current Trends International Semantic Web Conference (ISWC-05), pp. 506-521, 2005.
in Theory and Practice of Computer Science, Invited Talk, 2006. [39] D. McGuiness, R. Fikes, J. Rice, S. Wilder. An Environment for Merging
[15] G. Flouris, D. Plexousakis, G. Antoniou. Generalizing the AGM Postu- and Testing Large Ontologies. In Proceedings of the 7th International
lates: Preliminary Results and Applications. In Proceedings of the 10th Conference on Principles of Knowledge Representation and Reasoning
International Workshop on Non-Monotonic Reasoning, 2004. (KR-00), 2000.
[16] G. Flouris, D. Plexousakis, G. Antoniou. On Applying the AGM Theory [40] T. Meyer, K. Lee, R. Booth. Knowledge Integration for Description
to DLs and OWL. In Proceedings of the 4th International Semantic Web Logics. In Proceedings of the 7th International Symposium on Logical
Conference (ISWC-05), pp. 216-231, 2005. Formalizations of Commonsense Reasoning, 2005.
[17] N. Foo. Ontology Revision. In G. Ellis, R. Levinson, W. Rich, J. Sowa [41] G.A. Miller. WordNet: A Lexical Database for English. Communications
(eds). Proceedings of the 3rd International Conference on Conceptual of the ACM (CACM), 38(11):39-41, 1995.
Structures (ICCS-95), Lecture Notes in Artificial Intelligence (LNAI), [42] P. Mitra, N.F. Noy, A.R. Jaiswal. OMEN: A Probabilistic Ontology
Volume 954, Springer-Verlag, pp. 16-31, 1995. Mapping Tool. In Proceedings of the 4th International Semantic Web
[18] E. Franconi, F. Grandi, F. Mandreoli. A Semantic Approach for Schema Conference (ISWC-05), pp. 537-547, 2005.
Evolution and Versioning in Object-Oriented Databases. In Proceedings [43] N.F. Noy, R. Fergerson, M. Musen. The Knowledge Model of Protégé-
of the 6th International Conference on Rules and Objects in Databases, 2000: Combining Interoperability and Flexibility. In Proceedings of the
2000. 12th International Conference on Knowledge Engineering and Knowl-
[19] P. Gärdenfors. Belief Revision: An Introduction. In P. Gärdenfors (ed). edge Management: Methods, Models, and Tools (EKAW-00), pp. 1732,
Belief Revision, pp. 1-20, Cambridge University Press, 1992. 2000.
[20] P. Gärdenfors. The Dynamics of Belief Systems: Foundations Versus [44] N.F. Noy, M. Klein. Ontology Evolution: Not the Same as Schema
Coherence Theories. Revue Internationale de Philosophie 44, 1992. Evolution. Knowledge and Information Systems, 6(4):428-440, 2004.
[21] T.R. Gruber. A Translation Approach to Portable Ontology Specifica- [45] N.F. Noy, S. Kunnatur, M. Klein, M.A. Musen. Tracking Changes
tions. Knowledge Acquisition, 5(2):199-220, 1993. During Ontology Evolution. In Proceedings of the 3rd International
[22] P. Haase, L. Stojanovic. Consistent Evolution of OWL Ontologies. In Semantic Web Conference (ISWC-04), 2004.
Proceedings of the 2nd European Semantic Web Conference (ESWC- [46] N.F. Noy, M.A. Musen. Algorithm and Tool for Automated Ontology
05), 2005. Merging and Alignment. In Proceedings of the 17th National Conference
[23] P. Haase, Y. Sure. D3.1.1.b State of the Art on Ontology Evolution. on Artificial Intelligence (AAAI-00), 2000.
SEKT Deliverable, 2004. [47] N.F. Noy, M.A. Musen. An Algorithm for Merging and Aligning On-
[24] J. Heflin, J. Hendler. Dynamic Ontologies on the Web. In Proceedings tologies: Automation and Tool Support. In Proceedings of the Workshop
of the 17th National Conference on Artificial Intelligence, 2000. on Ontology Management at 16th National Conference on Artificial
[25] J. Heflin, J. Hendler, S. Luke. Coping with Changing Ontologies in a Intelligence (AAAI-99), 1999.
Distributed Environment. In Proceedings of the Workshop on Ontology [48] N.F. Noy, M.A. Musen. SMART: Automated Support for Ontology
Management of the 16th National Conference on Artificial Intelligence Merging and Alignment. In Proceedings of the 12th Workshop on
(AAAI-99), WS-99-13, AAAI Press, pp. 74-79, 1999. Knowledge Acquisition, Modeling and Management, 1999.
[49] R. Pan, Z. Ding, Y. Yu, Y. Peng. A Bayesian Network Approach to
Ontology Mapping. In Proceedings of the 4th International Semantic
Web Conference (ISWC-05), pp. 563-577, 2005.
[50] R.J. Peters, T. Ozsu. An Axiomatic Model of Dynamic Schema Evolu-
tion in Objectbase Systems. ACM Transactions on Database Systems,
22(1):75-114, 1997.
[51] H.S. Pinto, A. Gomez-Perez, J.P. Martins. Some Issues on Ontology
Integration. In Proceedings of the Workshop on Ontologies and Problem-
Solving Methods (KRR5), 1999.
[52] P. Plessers, O. De Troyer. Ontology Change Detection Using a Version
TABLE I
Log. In Proceedings of the 4th International Semantic Web Conference,
2005. S UMMARY OF THE VARIOUS SUBFIELDS OF ONTOLOGY CHANGE
[53] G. Stoilos, G. Stamou, S. Kollias. A String Metric for Ontology Align-
ment. In Proceedings of the 4th International Semantic Web Conference
(ISWC-05), pp. 624-637, 2005. Ontology Purpose: Heterogeneity resolution, interoperability
[54] L. Stojanovic, A. Maedche, B. Motik, N. Stojanovic. User-driven On- Mapping Input: Two (heterogeneous) ontologies
tology Evolution Management. In Proceedings of the 13th International Output: A mapping between signatures
Conference on Knowledge Engineering and Knowledge Management, Properties: Output identifies related signature entities
2002. Ontology Purpose: Heterogeneity resolution, interoperability
[55] L. Stojanovic, A. Maedche, N. Stojanovic, R. Studer. Ontology Evolu- Morphism Input: Two (heterogeneous) ontologies
tion as Reconfiguration-Design Problem Solving. In Proceedings of the Output: Mappings between signatures and axioms
2nd International Conference on Knowledge Capture, 2003. Properties: Output identifies related signature entities
[56] H. Stuckenschmidt, M. Klein. Integrity and Change in Modular On- and axioms
tologies. In Proceedings of the 18th International Joint Conference on
Ontology Purpose: Heterogeneity resolution, interoperability
Artificial Intelligence (IJCAI-03), 2003.
Alignment Input: Two (heterogeneous) ontologies
[57] G. Stumme, A. Maedche. Ontology Merging for Federated Ontologies
Output: A relation between signatures
on the Semantic Web. In Proceedings of the International Workshop on
Properties: Output identifies related signature entities
Foundations of Models for Information Integration (FMII-01), 2002.
[58] R. Wassermann. Revising Concepts. In Proceedings of the 5th Workshop Ontology Purpose: Heterogeneity resolution, interoperability
on Logic, Language, Information and Communication (WoLLIC-98), Articulation Input: Two (heterogeneous) ontologies
1998. Output: An intermediate ontology and mappings be-
[59] R. Wassermann, E. Fermé. A Note on Prototype Revision. In Spinning tween the signatures of the intermediate on-
Ideas, 1999. tology and each source
[60] Z. Zhang, L. Zhang, C.X. Lin, Y. Zhao, Y. Yu. Data Migration for Properties: Output is equivalent to a relation identifying
Ontology Evolution. In Poster Proceedings of the 2nd International related signature entities
Semantic Web Conference (ISWC-03), 2003. Ontology Purpose: Use a different representation language
Translation Input: Ontology and target representation language
(first Output: Ontology expressed in the target language
reading) Properties: Produces an equivalent ontology, if possible
Ontology Purpose: Implementation of a signature mapping
Translation Input: An ontology and a mapping
(second Output: An ontology
reading) Properties: Implements the mapping
Ontology Purpose: Apply changes (domain/conceptualization)
Evolution Input: Ontology and change operation(s)
Output: An ontology
Properties: Implements change(s) to the source ontology
Ontology Purpose: Transparent access to different versions
Versioning Input: Different versions of an ontology
Output: A versioning system
Properties: Version ids identify versions; transparent ac-
cess to versions; compatibility determination
Ontology Purpose: Fuse ontologies; similar domains
Integration Input: Two ontologies (covering similar domains)
Output: An ontology
Properties: Fuses knowledge to cover a broader domain
Ontology Purpose: Fuse ontologies; identical domains
Merging Input: Two ontologies (covering identical domains)
Output: An ontology
Properties: Fuses knowledge to describe the domain
more accurately