=Paper=
{{Paper
|id=Vol-370/paper-1
|storemode=property
|title=Emergent Semantics and Cooperation in Open Systems
|pdfUrl=https://ceur-ws.org/Vol-370/keynote.pdf
|volume=Vol-370
|authors=Keynote statement by
}}
==Emergent Semantics and Cooperation in Open Systems==
https://ceur-ws.org/Vol-370/keynote.pdf
Emergent Semantics and Cooperation in Open Systems
[Keynote Abstract]
Tiziana Catarci
Dipartimento di Informatica e Sistemistica “A.Ruberti”
SAPIENZA - Università di Roma
Via Ariosto 25 - 00185 Roma, Italy
catarci@dis.uniroma1.it
Information systems of every organization (ranging from semantics to human cognition, some researchers have pro-
large companies to individual entities) have to handle a vari- posed to use an explicit, shared reference system for relating
ety of information sources, from proprietary ones to informa- sets of symbols. Ontologies serve this purpose: they consist
tion publicly available in web services worldwide. Grasping of explicit, partial definitions of the intended meaning of
relevant information wherever it may be and exchanging in- symbols for a domain of discourse. Unfortunately, build-
formation with all potential partners has become an essential ing shared ontologies is a complex process and top-down
challenge. Basically, information sharing, rather than infor- ontology design, even when done collaboratively, is known
mation processing, is IT’s primary goal in the 21st century. not to scale well. Moreover, ontologies are not enough to
The key point is that now information has to be sharable in achieve semantic interoperability. For instance, ontologies
an open environment, where interacting peers do not neces- are forms of “a-priori” agreements on concepts, and there-
sarily have a common understanding of the world at hand, fore, their use is insufficient in ad-hoc and dynamic situa-
as used to be the case in traditional enterprise information tions where the interacting parties did not anticipate all the
systems. Lack of common background generates the need for interpretations and where “on-the-fly” integration must be
explicit guidance in understanding the exact meaning of the performed [4]. Indeed, emergent semantics is a global state
data, i.e., their semantics. Data semantics is more and more that should result from the dynamics of local interactions,
context- and time-dependent, and cannot be fixed once and without any predefined agreement. Such a state cannot be
for all at design time. Identifying emerging relationships predicted from individual behaviors, nevertheless single in-
among previously unrelated information items (e.g., dur- teracting peers should be able to analyze feedback from the
ing data interchange) may dramatically increase their value. overall network and infer from such a feedback the reliability
Such relationships are the basic ingredients for semantic in- of shared context. Given their characteristics, emergent se-
teroperability that is viewed as an emergent phenomenon mantics systems are typically peer-to-peer and implemented
constructed incrementally, and its state at any given point on top of so-called semantic overlay networks [1]. Research
in time depends on the frequency, the quality and the effi- on such systems is going on, still many open problems ex-
ciency with which negotiations can be conducted to reach ist, e.g., global semantic integrity and global consensus; ef-
agreements on common interpretations within the context ficiency and scalability; trust, quality, and reputation; auto-
of a given task. This type of semantic interoperability is matic construction of local consensus; resource location and
referred to as “emergent semantics” [3, 5]. identification; uncertain, imprecise, inconsistent, and incom-
plete information.
Software agents have various mechanisms at their disposal
for establishing relationships between internal symbols and An example of systems dealing with semantic interoperabil-
external meaning. In many cases, humans are responsible ity in dynamic open environments, i.e. emergent semantics,
for providing the initial semantics. In the simplest case, is Esteem (Emergent Semantics and cooperaTion in multi-
the natural language vocabulary is used for the local sym- knowledgE EnvironMents) [6].
bols and their relationship with the definition of the notion
concerned is left implicit. Often, the hidden assumption
is that the local symbol meaning is identified through hu- The Esteem approach proposes a comprehensive framework
man cognition. In order to address some of the problems and platform for data and service discovery in P2P sys-
arising when leaving interpretation of the symbol implicit tems, with advanced solutions for trust and quality-based
data management, P2P infrastructure definition, query pro-
cessing and dynamic service discovery in a context-aware
scenario. The system allows one to access data and ser-
vices in a simple and effective way, by querying informa-
tion sources that are similar to the user’s interests. Com-
mon interests identify semantic communities, which repre-
sent semantic affinity between peers emerging in a dynamic
and heterogeneous environment. Data and service discov-
ery is performed inside the borders of such communities.
User Assistant
Context Context Query
Matching Manager Manager
Data & service
discovery
Quality P2P Mapping
Service Manager Manager
Matching
Semantic
matchmaking
Ontology Semantic
Community Semantic Semantic
Network & overlay Matching community & routing
Membership Neighbor Routing
Global Semantic Preferential
Overlay Overlay Link
Figure 1: The reference architecture of an Esteem peer
Semantic communities are created and updated in an auto- global overlay) in order to enforce effective data and service
matic way by collecting information sources whose contents sharing. In this respect, the probe/search mechanism is used
present high similarity. A threshold-based mechanism al- to characterize:
lows to establish internal cohesion of community contents,
enabling peer aggregation apart of their terminological dif-
ferences. Semantic communities do not constrain partici- • the discovery phase, based on ontology matching, where
pants to adhere to a global ontology, but compare reference probe queries are defined to identify the peers that are
ontologies used by information sources belonging to the same capable of providing relevant knowledge with respect
community. In the Esteem system context-aware data and to a given topic of interest;
service selection excludes from search results resources that
• the sharing phase, based on P2P mapping definition,
are not accessible from the particular user context. The
where standard search queries are issued to point-to-
Esteem system is also in charge of protecting the users from
point interact with a previously discovered peer for
retrieving data and services from untrustworthy information
actual data acquisition and/or service invocation.
sources.
In the discovery phase, the joined semantic communities are
Esteem relies on an overlay P2P network where i) seman- exploited by a requesting peer for selecting the probe query
tic communities are defined to aggregate peers with simi- recipients with the aim of choosing those communities and
lar interests and ii) a probe/search mechanism is adopted peers that are most likely to provide relevant results ac-
to enforce data and service discovery/sharing. An Esteem cording to the query target. In this context, the semantic
semantic community sc is defined as a pair of the form matchmaker is invoked to evaluate the relevance of a com-
sc = hCID, M i where CID is the unique Community Identi- munity with respect to a probe query by comparing the com-
fier that characterizes the community sc and M is the Man- munity manifesto against the query content. By collecting
ifesto, that is the community ontology that describes the probe query replies, a peer evaluates the results and decides
common interpretation (i.e., perspective) of the community whether to perform the sharing phase by directly interact-
interests. In Esteem, a semantic community is autonomously ing with the most interesting peers that provided a reply
emerging, in that it originates from a proposal of a commu- through appropriate search queries with the aim of access-
nity founder (i.e., a peer) which initiates the community ing their data and services.
formation through dissemination of an advertisement mes-
sage that contains CID and M of the emerging community.
Each receiving peer pi autonomously decides whether to join An Esteem peer is equipped with (all or a subset of): (i) a
the community on the basis of its level of interest in the re- semantic description of shared data and services, to properly
ceived manifesto M . Such a level of interest is computed identify its interests, expressed through ontologies; (ii) the
by invoking an ontology-based semantic matchmaker and representation of context(s) from which the peer accesses
by evaluating the semantic affinity between M and the peer data and invokes services; (iii) the representation of quality
ontology of pi . Furthermore, communities are exploited as and trust metadata attached to its data and services (quality
a semantic overlay on top of the basic P2P overlay (i.e., the profile). When joining semantic communities that share its
interests, the peer also maintains information about joined [4] A. Ouksel. In-Context Peer-to-Peer Information
communities. The Esteem architecture addresses the main Filtering on the Web. SIGMOD Record, 32(3), 2003.
requirements of a peer involved in P2P semantic coopera- [5] P. Cudrè-Mauroux et al. Viewpoints on Emergent
tion. As shown in Figure 1, the main components are: Semantics. Journal on Data Semantics (JoDS), VI,
2006.
• Network & overlay. It is responsible for managing [6] The ESTEEM Project Web Site. Emergent Semantics
the peer connectivity and for handling incoming and and cooperaTion in multi-knowledgE EnvironMent.
outgoing messages. From the network point of view, http://www.dis.uniroma1.it/∼esteem/.
the Esteem P2P infrastructure is organized in semantic
overlays featuring the semantic communities. In this
respect, the network & overlay component is respon-
sible for maintaining the overlays and the associated
peer communications.
• Semantic community & routing. It is responsible for
managing the peer participation in semantic communi-
ties and for discovering the semantic neighborhoods of
a peer. Furthermore, this component is responsible for
providing a semantic routing mechanism to effectively
enforce query propagation.
• Semantic matchmaking. It is responsible for providing
semantic affinity evaluation when comparing different
peers’ ontological descriptions. This component is in-
voked by a peer during the discovery phase to identify
peers that are capable of providing matching resources
(i.e., data, service, context) w.r.t. a given target re-
quest. Different techniques are provided by the se-
mantic matchmaking component according to the type
of matching resource that is specified in the request.
In particular, ontology, service, and context matching
techniques are provided by the semantic matchmaker.
• Data & service discovery. It is responsible for inter-
acting with the user and for satisfying its discovery
requests. In particular, this component provides the
functionalities for context and quality/trust manage-
ment. Furthermore, discovery and sharing function-
alities are also addressed in this component through
query/answer and P2P mapping management.
Moreover, the system supports the (human) user with an in-
tuitive Web interface that assists him/her in joining the se-
mantic communities that share his/her interests and in iden-
tifying data and services he/she is looking for. The Esteem
system has been validated with a set of specialists from the
health-care domain both to collect system requirements in
the first stage of the project and to test the system through
a think aloud evaluation technique [2], getting satisfactory
results.
1. REFERENCES
[1] A. Crespo and H. Garcia-Molina. Semantic Overlay
Networks for P2P Systems. In Proc. of the 3rd Int.
Workshop on Agents and Peer-to-Peer Computing
(AP2PC 2004), New York, NY, USA, 2004.
[2] A. Dix, J. Finlay, G. Abowd, and R. Beale.
Human-Computer Interaction (3rd Edition).
Prentice-Hall, 2003.
[3] K. Aberer et al. Emergent Semantics Principles and
Issues. In Proc. of the 9th Int. Conference on Database
Systems for Advances Applications - DASFAA 2004,
Jeju Island, Korea, 2004.