A Domain-Specific Formal Ontology for
Archaeological Knowledge Sharing and Reusing
Chunxia Zhang, Cungen Cao, Fang Gu, and Jinxin Si
Key Laboratory of Intelligent Information Processing, Institute of Computing Technology,
Chinese Academy of Sciences, Beijing 100080, China
{cxzhang, cgcao, fgu, jxsi}@ict.ac.cn
Abstract. Inherent heterogeneity and distribution of knowledge strongly prevent
knowledge from sharing and reusing among different agents and across different
domains; formal ontologies have been viewed as a promising means to tackle
this problem. In this paper, we present a domain-specific formal ontology for
archaeological knowledge sharing and reusing. The ontology consists of three
major parts: archaeological categories, their relationships and axioms. The ontology not only captures the semantics of archaeological knowledge, but also
provides archaeology with an explicit and formal specification of a shared conceptualization, thus making archaeological knowledge shareable and reusable
across humans and machines in a structured fashion. As an application of the
ontology, we have developed an ontology-driven approach to knowledge acquisition from archaeological text.
1 Introduction
With the widespread use of computers and the World Wide Web, inherent heterogeneity and distribution of domain knowledge strongly prevent it from sharing and reusing across different agents and domains. This happens to every domain, and archaeology is no exception! Archaeological knowledge is an important part of human
knowledge, and is necessary in many knowledge-intensive applications, such as
knowledge-based systems, information retrieval and intelligent tutoring for archaeology. However, it lacks a tool for giving itself a formal, clear and declarative description for communicating, sharing and reusing archaeological knowledge among
different humans and software entities.
Formal ontologies have been viewed as a promising means to tackle this problem.
However, like formal methods in software development, they are not often used directly in modeling applications because of their formalities and abstractness. We
1
This work is supported by a grant from the Chinese Academy of Sciences (#2000-4010), a
grant from the Foundation of Chinese Natural Sciences (#20010010-A), and a grant from the
Ministry of Science and Technology (#2001CCA03000).
D. Karagiannis and U. Reimer (Eds.): PAKM 2002, LNAI 2569, pp. 213-225, 2002.
© Springer-Verlag Berlin Heidelberg 2002
214
C. Zhang et al.
would argue that domain-specific ontologies are more useful and theoretically valuable.
A domain-specific ontology of archaeology in this paper captures the semantics of
archaeological knowledge; gives the domain an explicit and formal specification of a
shared conceptualization; and makes it shareable and reusable for agents in a structured way [2, 8, 14]. It is also a basis for an ontology-driven approach to knowledge
acquisition from text [5], and serves as a guideline for defining a natural language
interface for querying archaeological knowledge [9].
The rest of the paper is organized as follows. Section 2 gives a few basic definitions
in our domain-specific ontology of archaeology. Section 3 discusses principles of
domain ontology design, content and representation. Comparison between our ontology and the object-oriented concepts reference model of the ICOM-CIDOC (International Committee for Documentation of the International Council of Museums) is
presented in section 4. Section 5 concludes the paper and raises a few problems on our
research agenda.
2 Preliminaries
A domain-specific ontology of archaeology is an explicit and formal specification of a
shared conceptualization of archaeology. Here, conceptualization refers to an abstract
model of some phenomenon in the world by having identified the relevant categories
of that phenomenon; explicit means that the type of categories and the constraints on
their use are explicitly defined; formal refers to that the ontology should be machinereadable; and shared means that the ontology captures consensual knowledge (it is not
private to some individual, but accepted by a group) [7, 17]. The domain ontology is
composed of categories, their relationships and formal axioms that constrain the interpretation and use of these terms, as shown in the upper part of Fig.1.
Dom ain-Specific
O ntology of Archaeology
Category
Structure
Axioms
Inner-Categories
Inner-Categories
Axiom
Axiomss
Slot
Slot Axiom
Axiomss
of
of Categories
Categories
Category
M
Mem
embership
bership Axiom
Axiomss
of
of Categories
Categories
Attributes
Relationships
Axiom s
Archaeological Knowledge
Fig. 1. Components of Our Domain-Specific Ontology of Archaeology
Definition 1. Given a category C1 in a space of archaeological categories, the individual space of C1 is the set of all individuals of C1, written as ind(C1).
A Domain-Specific Formal Ontology for Archaeological Knowledge Sharing
215
For example, the category Person consists of all human individuals – living or
dead; that is, ind(Person) is the set of all human individuals. As a point in a space of
categories, a specific category is related to others in one way or another. In the following, we define a few useful relationships between categories.
Definition 2.Given two categories C1 and C2, if ind(C2)±ind(C1), we say that C2 is a
specialization of C1, written as is-spec(C2, C1). In other words, C1 is a super-category
of C2, and C2 is a sub-category of C1.
Definition 3.Given two categories C1 and C2, we define a mapping q as
q: ind (C1) “ ind (C2)
"a³ind (C1), if q(a) is part of a; q(a)³ind (C2); and q is a surjection; then we say
that C2 is a part-category of C1, and C1 is a whole-category of C2, written as part-of
(C2, C1).
To articulate the part-of relationship in the domain of archaeology, we reconsider a
new classification of part-of. In the classification of part-of given by [1], we find that
the object/stuff relationship is not adequate of specifying the part-of relationship in
archaeology. The reason is that stuff is an attribute attached to an object, and it has no
independent existence. In this sense it is unreasonable to regard stuff as a category.
We summarize our classes of part-of relationships of archaeological objects as
follows:
1. Integral-object /Component. For example, “The wall is part of a carcass”.
2. Area/Place. For example, “A tomb is part of a graveyard”.
3. Activity/Feature. For example, “The design is part of a piece of pottery”.
4. Collection/Member. For example, “An artifact is part of a complete set of artifacts”.
5. Mass/Portion. For example, “A block of accumulation is part of an accumulation layer”.
3 A Domain-Specific Formal Ontology of Archaeology
Before building a domain-specific ontology of archaeology, we must consider some
basic problems: What are principles of constructing it? How to represent it? What is
the specific content of it? What are its applications? Etc.
3.1 Design Criteria
The principles of constructing ontologies have been expounded in [7, 10, 13, 14]. For
the main purpose of sharing and reusing knowledge, designing a domain-specific
ontology of archaeology should comply with the following criteria.
- Sharability and reusability. The domain ontology provides an explicit and formal
conceptual specification and taxonomic structure of archaeological objects, so exchanging and communication between different agents will become easier and convenient. In addition, the knowledge engineers who build other domain ontologies
can share and reuse our archaeology ontology.
216
C. Zhang et al.
- Clarity and objectivity. The meaning of each term is provided with objective and
natural language definitions.
- Completeness. Definitions of categories are expressed in terms of necessary and
sufficient conditions.
- Transformation. The domain ontology is also thought of as an object-oriented semantic model, if its axioms are ignored. Therefore it can be easily converted into
other object-oriented models.
- Readability. Archaeological knowledge representation is natural and expressive to
archaeological domain experts.
- Appellation Standardization. We first adopt the appellations of categories, attributes
and relationships of the Standard in Archaeology [18]. If these terms do not exist in
the Standard [19], they are named by consulting archaeologists.
3.2 Ontology Representation
First of all, we must emphasize that an ontology has nothing to do with any formal
language. But for communicating ontologies with other people, we have to specify
them in a certain formal language. Our ontology language is a combination of a frame
formalism and first-order predicate calculus. It uses frame mechanism from the Generic Frame Protocol [6] and Loom [3, 20]. Fig.2 illustrates a frame-based schema for
representing categories. Axioms are first-order formulae, representing a first-order
theory on categories and their relationships.
The reasons that we don’t adopt other representation approaches (such as description logics, full first-order logics and logic programming) are as follows. Description
logic describes the hyponyms among concepts through the containing relation between
different individual sets. However, it can’t depict the meronyms, and doesn’t represent
the logical relation among concepts. In addition, knowledge about a concept or an
individual, represented in full first-order logic and logic programming, is partitioned
into a large number of independent pieces of propositions in the knowledge base,
while the frame mechanism organizes this kind of knowledge into a frame. Furthermore, the frame representation is equivalent to the first-order logic representation,
which makes it convenient knowledge inference. Our ontology language not only
denotes the hyponyms and the meronyms, but also logical relation among categories.
In our language, a category specification has three parts. The category header begins with the keyword defcategory, followed by the name of the category to be defined.
The category body consists of a (unordered) list of slots. A slot is classified as an
attribute or a relationship. It may have a number of facets. A slot facet provides additional information for a slot. Typically, we have
- Type: The value type of the slot. This facet is mandatory for all slots.
- Domain: The set of complete values the slot can assume
- Default: The default value of the slot
- Unit: The unit of the slot value
- Synonym: The synonyms of the slot
A Domain-Specific Formal Ontology for Archaeological Knowledge Sharing
217
- Parasynonym: The parasynonyms of the slot
- Antonym: The antonyms of the slot
- Reverse: The reverse slot of the slot. For example, the reverse of the slot havemember is is-member-of.
- Property: The properties which the slot has.
- Facet: The domain-specific facets that the slot is subject to
- Related slots: The relevant slots related to the slot
- Comment: The informal comment on the slot
d efcateg o ry <category-nam e>[<related-categories>]
{
{<definition-of-slots>}
{<inner-category-ax iom s>}
}
<related-categories>::= S u p er-categ o ries <sequence-of-super-category>
[; P art-categ o ries <sequence-of-part-category>]
| P art-categ o ries <sequence-of-part-category>
[; S u p er-categ orie s <sequence-of-super-category>]
<defin ition-of-slots>::==<type-of-slot><slot-nam e>
{
:typ e <type-slot-va lue>
[:d o m ain <com plete-value-dom ain>]
[:d efault<default-value >]
[:u n it<units >]
[:syn o n ym <sequence-of-synon ym >]
[:p arasyn o n ym <sequence-of-parasynon ym >]
[:an ton ym <sequence-of-anton ym >]
[:rev erse<reverse-slots>]
[:p ro p erty<general-properties><specific-properties>]
[:facet<sequence-of-necessary-facet-users-define>]
[:related -slots<other-slots-related-w ith-the-slot>]
[:co m m e nt<inform al-com m ents>
……
<general-properties >::==< R eflexive | Irreflexive | A nti-reflexive | S ym m etric | A nti-s ym m etric|
A sym m etric | T ransitive | A nti-transitive | N on-transitive >
<specific-properties>::= =<tim e-varying |… … >
<inne r-category-ax iom s>::= =<A xiom s-of-definition-of-category>|< A xiom s-of-properties-ofcategory>|< A xiom s-of-slots> |<A xiom s-of-relationships-am ong-slots>
<A xiom s-of-definition-of-category>:: = =<first-order w e ll-form ed form ula>
<A xiom s-of-properties-of-category>:: == <first-order w ell-form ed form ula>
<A xiom s-of-slots>:: ==<first-order w e ll-form ed form ula>
<A xiom s-of-relationships-am ong-slots>:: == <first-order w e ll-form ed form ula>
}
Fig. 2. Category Representation (Keywords are in bold face)
The axioms constrain the interpretation of the slots and the relationships between
categories.
3.3 Categories and Their Backbone Structure
We identify main categories and their relationships in the domain of archaeology in
appropriate terms, and combine top-down and bottom-up strategies to capture other
categories. Moreover, we give precise meaningful models for the categories and their
218
C. Zhang et al.
relationships at the knowledge level. So far, we have built more than 120 archaeologyspecific categories. Fig.3 depicts the category Archaeological-Culture.
defcategory Archaeological-Culture Part-Categories: Cognition, Neo-Confucian-Orthodoxy, ArchaeologicalSite, and Archaeological-Relics
{
Attribute: Period
:Type string
:Domain Paleolithic-Period, and Neolithic-Period, and ……
:Synonyms Age
:Comment An interval of time characterized by the prevalence of a specified culture
Attribute: Upper-Limit-Year
:Type string
:Facet Dating-Method, and revision
:Comment The upper limit year of period of a specified culture
Attribute: Lower-Limit-Year
:Type string
:Facet Dating-Method, and revision
:Comment The lower limit year of period of a specified culture
Attribute: Year
:Type string
:Comment The fuzzy year of period of a specified culture
Relation: Developed-from
:Type string
:Comment Value domain is cultures which the specified culture developed from.
Relation: Evolve-to
:Type string
:Comment Value domain is cultures which the specified culture evolved to.
……
<Axioms of Properties of Category>
Archaeological-culture (a)→(∃b)(Period(a, b)∈Framea ∧b∈{sequences of period } )
<Axioms of Slots>
Archaeological-culture(a) ∧ Geological-Ages(a, b)→((∀c)Geological-Ages(a,c) →Equal(b,c))
<Axioms of Relationships of Slots>
Archaeological-Culture (a) ∧ Upper-Limit-Year (a, b1) ∈Framea ∧ Lower-Limit-Year (a, b2)
∈Framea →Earlier-Than (b1, b2)
……
}
Fig. 3. The Category Archaeological-Culture
There are a variety of relationships among categories. We mainly investigate two
kinds of the most fundamental relationships in the ontology: is-spec and the part-of
relationships. With these two relationships, we define the backbone structure of the
formalized archaeological categories, as depicted in Fig.4. In the figure, because of the
limitation of space, some branches are folded and those who are interested in the
backbone structure may contact with the first author to get a full copy thereof.
3.4 Archaeological Knowledge Representation
We use Framea to denote the frame for object a. Framea consists of a set of slotvalue pairs with optional facet-value pairs, i.e. Framea= {(slot, slot-value [,facet-1,
facet-1-value, … , facet-n, facet-n-value])}.
A Domain-Specific Formal Ontology for Archaeological Knowledge Sharing
219
For representing axioms, we define predicates of the form Slot(a, slot-value,
facet-1, facet-1-value, …, facet-n, facet-n-value) to denote the tuple of the
Framea. And predicate ½Slot(a) denotes that the object a doesn’t have the slot,
Predicate ½Slot(a, b) represents that the slot value of the object a is not b. For example, Upper-Limit-Year(Yangshao-culture, about B.C. 5000 Year, DatingMethod, Radiocarbon-Dating, Revision, Yes)³ FrameYangshao-culture.
Village
Shell-Mound-Site
City-Site
Is-spec Relationship
Part-of Relationship
Archaeological-Culture
Dwelling-Floor
Cooking-Stove
Post-Hole
Slope
Gateway
Fireplace
Export
Ground
Furnace
Dwelling-Vault
Well
Pit ……
Penstock
Carcass
Cognition ……
Entrenchment
Petrogram
Neo-Confucian-orthodoxy ……
Wall
Fence
Living-Site
DwellingBench-Table
Bounding-Wall
ArchaeoloSite
Roof
Square
gical-Site
Tomb …… Production-Site
Top-of-a-KitHole
chen-Range
Ecofact ……
Relics
Channel
Ash-Stack
Earth-Table
Relics-worked-on
Platform-ofPalace
Artifact
Rammed-Earth
Temple
Productive-Tool
Altar
Living-Appliance
……
Burial-Articles
Pottery ……
Polished-Stoneware
Stoneware
Chipped-Stoneware
Metalwork ……
……
Fig. 4. The Backbone Structure of Archaeological Categories
3.5 Semantic Axioms
Axioms are a crucial part of our domain ontology, and they constrain the interpretation
of ontological categories and their relationships and contents. In our ontology, axioms
are classified into three types according to the objects which they describe.
220
C. Zhang et al.
3.5.1 Membership Axioms of Categories
For each category, there are some axioms for determining whether an instance or
individual belongs to it. Formulating membership axioms for a category is difficult
since it requires a comprehensive understanding of archaeology and its domain knowledge.
How is a physical object judged
to be an archaeological-site or a
relic?
Axiom 1. For a physical object a,
Archaeological-site(a)↔ (∃b)( Location(a, b)∈Framea)
Axiom 2. For a physical object a,
Relics(a)↔(∀b)(¬(Location(a,b) ∈Framea))
Fig. 5. Illustrating Membership Axioms of Archaeological-Site and Relics
Fig.5 illustrates two membership axioms for Archaeological-Site and Relics
based on our understanding. Physical objects of archaeology fall into two categories
Archaeological-Site and Relics, and their difference is that the former demands the
location as its necessary elementary condition, while the identity of a piece of relics
is not necessarily related to a geo-location.
Based on the same reasoning, we have summarized numerous membership axioms
for other categories, some of which are shown as in Fig.6.
Axiom 3. Artifact (a) ↔Relic (a) ∧ ∃b ∃c ((Producer (a, b)∈Framea ) ∧ (Have-Use (a, c)
∈ Framea))
Axiom 4. Stone (a) ↔ Artifact (a) ∧ (Producing-Material (a, stone) ∈Framea)
Axiom 5. Ground-Stone (a) ↔ Stone (a) ∧(Producing-Method (a,ground) ∈Framea)
Axiom 6. Metalwork (a) ↔ Artifact (a) ∧ (Producing-Material (a, metal) ∈Framea)
Axiom 7. Ironware (a) ↔Metalwork (a) ∧ (Producing-Material (a, iron) ∈Framea)
Fig. 6. Membership Axioms of Artifact and Some of its Subcategories
Axiom 3 means that given an object a, a is a piece of Artifact if and only if there
exists a producer b of a, and a has a specific use (denoted as c). Other axioms have a
similar interpretation. From the axioms above, we can derive a lot of interesting theorems or propositions, and some are presented below.
Proposition 1. Given a physical object a, Ironware (a)“ Artifact (a).
Proof. Given a physical object a, if Ironware (a) holds, then we have Metalwork (a)
by axiom 7. Further, we derive Artifact (a) according to axiom 6.
Proposition 2. Given a physical object a, Ground-Stone (a) “ Artifact (a)
Proposition 3. Artifact (a) “ Relics (a)
Proposition 4. Artifact (a)“$b (Producer (a, b)³Framea)
Proposition 5. Artifact (a) “$c (Have-Use (a, c) ³Framea)
A Domain-Specific Formal Ontology for Archaeological Knowledge Sharing
221
3.5.2 Slot Axioms of Categories
In our ontology, slots are divided into two classes: attributes and relationships. As
illustration, we present a few axioms in the category Archaeological-Culture in Fig.
7. Axioms 8-10 are constraints on the two attributes period and founder and the relationship located-in, respectively. Axioms 11-14 are about some slots on time in Archaeological-Culture. Axiom 8 means that if an object a is an instance of Archaeological-Culture, then the value of the attribute period of a belongs to the set-ofsequences-of-periods such as Neolithic Age, Western Zhou and Qing.
From the axioms 8-14, we can derive the following propositions.
Proposition 6. Given an abstract object a, Archaeological-Culture(a)¾FoundTime(a, b)¾Upper-Limit-year(a, c1)¾Lower-Limit-Year(a, c2)“EarlierThan(c1,b)
Proposition 7. Given an abstract object a, Archaeological-Culture(a)¾Upper-LimitYear(a, c1) ¾Found-Time(a, b)¾ Excavation-Time(a, d) “ Earlier-Than(c1, d).
Proposition 8. Given an abstract object a, Archaeological-Culture(a)¾Lower-LimitYear(a, c2)¾Found-Time(a, b)¾ Excavation-Time(a, d) “ Earlier-Than(c2, d).
3.5.3 Inter-categories Axioms
Currently in our ontology, we identify two types of inter-category relationships: category-level and slot-level relationships. By a category-level relationship we mean that
the relationship does not refer to slots of the related categories. By a slot-level relationship, we mean that the relationship between categories is reflected by the relationship(s) of the slots in the relevant categories.
Axiom 8. Archaeological-Culture (a) →(∃b) ((Period (a, b) ∈Framea )∧b∈{sequences of period } )
Axiom 9. Archaeological-Culture (a) →(∃b) ((Founder (a, b) ∈Framea)∧Human-being(b))
Axiom 10. Archaeological-Culture (a) →(∃b) ((Located-in(a, b)∈Framea)∧Region(b))
Axiom 11. Archaeological-Culture (a) ∧ (Typical-Site (a, b1, b2,…,bn ) ∈Framea )→(Site (a, b1)
∈Framea )∧ (Site (a, b2 ) ∈Framea )∧…∧ (Site (a, bn) ∈Framea )
Axiom 12. Archaeological-Culture (a) ∧ (Upper-Limit-Year (a, c1) ∈Framea )∧ (Lower-Limit-Year
(a, c2) ∈Framea )→Earlier-Than (c1, c2)
Axiom 13. Archaeological-Culture (a) ∧(Period (a, b) ∈Framea )∧(Upper-Limit-Year (a, c1)
∈Framea )∧ (Lower-Limit-Year (a, c2) ∈Framea )→In (c1, b) ∧ In (c2, b)
Axiom 14. Archaeological-Culture (a) ∧ (Found-Time (a, b) ∈Framea )∧(Lower-Limit-Year (a, c2)
∈Framea )→Earlier-Than (c2, b)
Fig. 7. Slot Axioms of the Category Archaeological-Culture
Axiom 15. Relics (a) ↔(Artifact (a) ∧¬Ecofact (a) ∧¬Worked-on-Object (a))∨(¬Artifact (a)
∧Ecofact(a) ∧¬Worked-on-Object (a)) ∨ (¬Artifact (a) ∧¬Ecofact (a) ∧Worked-on-Object (a))
Axiom 16. Archaeological-Culture (a)→ ∃b (Archaeological-Site (b) ∧ Part-of (b, a))
Axiom 17. Archaeological-Culture (a) → ∃b (Artifact (b) ∧ Part-of (b, a))
Fig. 8. Axioms of Category-Level Relationships between Categories
222
C. Zhang et al.
Axioms 15-17 represent category-level relationships between categories. For instance, axiom 15 declares that an object is a piece of Relics if and only if it is an Artifact, an Ecofact, or a Worked-on-object, where an Artifact is man-made, an Ecofact is unworked or natural, and a worked-on-object is worked-on by human.
Axiom 18. Archaeological-Site (a) ∧ ∃b ((Found-Place (b, a) ∈Framea) ∧ Fossil (b))
→ Period (a, Paleolithic-Age)
Axiom 19. Archaeological-Site (a) ∧ ∃b (Found-Place (b, a) ∈Framea)∧ Stone-Tool
(b)∧ (Producing-Method (b, ground) ∈Framea) →¬Period (a, Paleolithic-Age)
Fig. 9. Axioms of Slot-Level Relationships between Categories
In Fig.9, axioms 18-19 are constraints on category-level relationships between
categories. For example, axiom 18 is a formulation of the fact that there is no enough
time to form Fossil after the Paleolithic-Age according to archaeological stratigraphy
and formation process of Fossil.
3.6 Knowledge Acquisition and Inference Based on the Domain-Specific Ontology
Many endeavors have been intended to acquire knowledge from text. We can acquire
about 90 to 95 percent of the domain knowledge in knowledge-based systems directly
from domain texts [4, 8, 11, 12]. The ontology-driven archaeological knowledge acquisition process has as following steps [16]. First, identify categories and individuals
in texts of archaeological knowledge; second, find relations among categories and
individuals according to definitions of categories, their backbone structure and axioms; third, extract the attribute values and the relation values of every categories and
individuals.
The domain-specific ontology of archaeology is also used for knowledge analysis,
and knowledge inference, and knowledge discovery. In particular, a question answering system with a natural language interface for inquiring about archaeological information has been developed [9]. Based on the axioms of the domain ontology, we can
find inconsistent knowledge, such as inaccurate attribute values and relation values of
categories and individuals, and false is-spec and part-of relations. In addition, given a
proposition or problem, we can deduce proposition’s correctness or answer the problem through knowledge searching and inference. The procedure is divided into 3 steps.
First, extract entities, attributes, relations, and interrogative. Second, search for these
terms in frame-based archaeological knowledge base. Thirdly, if they are found, then
give the answers, otherwise, resort to axioms to problem solving. If the problem involves only one category, then utilize slot axioms of the category, or else, use intercategory axioms.
A Domain-Specific Formal Ontology for Archaeological Knowledge Sharing
223
4 A Comparison with CRM
To our knowledge, however, there exists few domain-specific ontology of archaeology
so far. The project team of post-excavation on complex sites at the University of York
is developing an archaeological ontology to key aspects of the archaeological process.
The object-oriented conceptual reference model (i.e., CRM) of the ICOM-CIDOC
could more or less be viewed as a domain ontology, and it is intended to cover all
concepts relevant to museum documentation. The model was developed by the
CIDOC Documentation Standards Group from 1994 to 1999 [17]. The relationship isspec is considered in both our domain ontology and CRM.
The differences between our ontology of archaeology and CRM are:
- Our ontology of archaeology orients all objects in the generic domain of archaeology, while CRM is a description of the information categories used for developing records about the objects pertinent to museum documentation.
- CRM deals with three kinds of part-of relationships: integral-object /component,
area/place, and activity/feature. Our domain ontology introduces collection/member
and mass/portion part-of relationships in addition.
- The key difference is that formal axioms are an indispensable component in our
domain ontology (Axioms are used for constraining the interpretation of ontological
categories and their relationships and contents), while CRM doesn’t provide formal
meanings of the entities and properties.
- Sites and holes are classified as objects in our ontology, while they are considered
as physical features in CRM. We think it is unreasonable, for if post-holes are not
physical objects, then they can’t occupy space and posts couldn’t occupy the same
space, though post-holes are attached to posts.
5 Conclusion
Sharing and reusing knowledge has become a significant research topic in information
community with the widespread of Internet. In this paper, we discuss our current work
of constructing an archaeological knowledge base with ontological theory and engineering. Categories, their relationships, and constraint axioms are three factors of a
domain-specific ontology of archaeology. Since domain knowledge is described in a
formal, explicit and structured way to provide a commonly agreed understanding of
archaeology, it can be shared and reused across different agents and knowledge bases.
We adopt a combination of a frame formalism and first-order predicate calculus as the
ontology representation language. The language can describe the hyponyms, the
meronyms, logical relations about categories and individuals. Production rules has the
disadvantages: inefficient and inflexible, though its advantages include naturalness of
expression, modularity and restricted syntax. A major problem with semantic network
is that although the name of this language is semantic nets, there is not clear semantics
of the various network representations. And it doesn’t convey some kinds of facts.
224
C. Zhang et al.
Based on the domain ontology, we have developed an ontology-driven approach to
archaeological knowledge acquisition, and a question answering system with a natural
language interface for inquiring about archaeological information.
There are a few problems to be solved in our project: How to utilize axioms to discover unknown knowledge? How to find potential applications of ontology? How to
formalize our ontology development method so that more fundamental issues can be
tackled vigorously (e.g. consistency, redundancy, and completeness of ontological
axioms)?
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
A.Artale, E.Franconi, Nicola Guarion, and Luca Pazzi.: Part-whole Relations in Objectcentered Systems: an Overview. Data & Knowledge Engineering, Vol.20 (1996) 347-383
W.N.Borst.: Construction of Engineering Ontologies. PhD thesis, University of Twente,
Enschede (1997)
David Brill.: Loom Reference Manual Version 2.0.
http://www.cbl.leeds.ac.uk/nikos/tex2html/examples/shell/shell.html (1993)
Cungen Cao.: Medical Knowledge Acquisition from Encyclopedic Texts. Lectures in
Computer Science, Vol.2101 (2001) 268-271
Cungen Cao, Qiangze Feng, et.al.: Progress in the Development of National Knowledge
Knowledge Infrastructure. Journal of Computer Science and Technology, Vol.17
(2002)523-534
V.Chaudhri, A. Farquhar, R. Fikes, P. Karp, & J. Rice.: The Generic Frame Protocol
2.0,KSL-97-05 (1997) http://www.ksl.stanford.edu/KSL_Abstracts/ KSL-97- 05.html
Ying Ding, Dieter Fensel.: Ontology Library Systems: The Key to Successful Ontology
Re-use. http://www.semanticweb.org/SWWS/program/full/paper58.pdf. (2001)
D.Fensel. Ontologies: a Silver Bullet for Knowledge Management and Electronic Commerce. Springer-Verlag, Berlin (2001)
Qiangze Feng, Cungen Cao, et al..: A Uniform Human Knowledge Interface to the MultiDomain Knowledge Bases in the National Knowledge Infrastructure. To appear in The
Twenty-second SGAI International Conference on Knowledge Based Systems and Applied
Artificial Intelligence (2002)
Asunción Gómez Pérez, V.Richard Benjamins. Overview of Knowledge Sharing and
Reuse Components: Ontologies and Problem-Solving Methods. Proceedings of the JICAI99 workshop on Ontologies and Problem-Solving Methods (KRR5) Stockholm, Sweden,
August 2 (1999)
Ruqian Lu, Cungen Cao.: Towards Knowledge Acquisition from Domain Books. In:
Wielinga, B., Gaines, B., Schreiber, G., Vansomeren, M. (eds.): Current Trends in Knowledge Acquisition. Amsterdam: IOS Press (1990) 289-301
Ruqian.Lu, Cungen. Cao, Y.Chen, et al.: A PNLU Approach to ICAI System Generation.
Science in China (Series A), vol.38 (1996) 1-10
Asunción Gómez-Pérez.: Evaluation of Taxonomic Knowledge in Ontologies and knowledge Bases. Proceedings of the Knowledge Acquisition Workshop, KAW99 (1999)
Thomas R. Gruber. Toward Principles for the Design of Ontologies used for Knowledge
Sharing. Knowledge Acquisition, vol.5, no.2, (1993) 199-220
R.Studer, V.R. Benjamins, and D. Fensel.: Knowledge Engineering, Principles and Methods. Data and Knowledge Engineering, 25(1-2) (1998) 161-197
A Domain-Specific Formal Ontology for Archaeological Knowledge Sharing
225
16. Jinxin Si, Cungen Cao, Haitao Wang, Fang Gu, Qiangze Feng, Chunxia Zhang, et al.: An
Environment for Multi-Domain Ontology Development and Knowledge Acquisition. Proceedings of Engineering and Development of Cooperative Information System. (2002)104116
17. The ICCOM/CIDOC Documentation Standards Group. Definition of the CIDOC Objectoriented Conceptual Reference Model. http://cidoc.ics.forth.gr/docs/cidoc_crm_version
3.3.2.doc (2002)
18. The Core Data Standard for Archaeological Sites and Monuments. http://www.cidoc
.icom.org/archo.htm (1995)
19. Standards in Archaeology. http://ads.ahds.ac.uk/project/userinfo/standards.html (2002)
20. Loom 4.0 Release Notes. http://www.isi.edu/isd/LOOM/how-to-get.html (2000)