=Paper=
{{Paper
|id=Vol-263/paper-4
|storemode=property
|title=Emergent Communities for Semantic Collaboration in Multi-Knowledge Environments: Methods and Techniques
|pdfUrl=https://ceur-ws.org/Vol-263/paper4.pdf
|volume=Vol-263
|dblpUrl=https://dblp.org/rec/conf/caise/Montanelli06
}}
==Emergent Communities for Semantic Collaboration in Multi-Knowledge Environments: Methods and Techniques==
https://ceur-ws.org/Vol-263/paper4.pdf
1152 CAiSE'06 Doctoral Consortium
Emergent Communities for Semantic
Collaboration in Multi-Knowledge
Environments: Methods and Techniques
Ph.D. Thesis Abstract
Stefano Montanelli
Università degli Studi di Milano
DICo - Via Comelico, 39, 20135 Milano - Italy
montanelli@dico.unimi.it
Abstract. The need of sharing data and resources to foster semantic
collaboration is a key problem at the current stage of development of
open distributed systems. In this context, autonomous and self-organizing
communities of peers emerge by bringing together those peers that are in-
terested in similar topics and plan to strengthen their cooperations. The
Ph.D. thesis abstract illustrates a semantic handshake process based on
ontologies and ontology matching techniques to handle consensus nego-
tiation and peer community formation. Furthermore, we discuss the pos-
sible benefits deriving from semantic community adoption by describing
the community-aware query propagation strategy for effective distribution
of resource requests on a semantic basis within a committed community.
1 The research question of the thesis
The need of sharing data and resources to foster semantic collaboration is a key
problem at the current stage of development of open distributed systems, like
P2P networks, and semantic Grids [2]. In this context, the emergence of col-
laboration among peers requires dynamic capabilities of negotiating agreements
on common interpretations within the context of a given task. This is typical
for instance of peer-based systems, characterized by a set of independent peer
parties without prior reciprocal knowledge and no degree of relationship, that
dynamically need to cooperate by sharing their resources (e.g., data, documents,
services). These collaboration scenarios are multi-knowledge, in that no central-
ized authorities are defined to manage a comprehensive view of the resources
shared by all the nodes in the system, due to the high dynamism and vari-
ability of collaboration and sharing requirements. On the opposite, each peer is
responsible of providing the knowledge description of the resources to be shared
through its own ontology. In order to facilitate resource discovery in such highly
dynamic and open contexts, the formation of autonomous and self-organizing
communities of parties poses new issues to be investigated and some work in
this direction has been appeared in the literature [1, 3, 8, 9]. Communities aim to
handle the problem of high network traffic due to single-peer interactions and to
CAiSE'06 DC 1153
provide a coordination mechanism for processing and forwarding resource queries
on a semantic basis, by exploiting available ontologies describing resources to be
shared. In this respect, the intrinsically open nature of P2P systems, and thus
of the communities, poses serious issues regarding the maintenance of the com-
munities and requires policies and mechanisms to specify the rules regulating
resource sharing and the conditions under which a peer is available to process
incoming queries.
With respect to this scenario, the Ph.D. thesis will be devoted to investigate
two main issues: i) the development of consensus-driven techniques which ex-
ploit ontological resource descriptions and ontology matching in order to form,
maintain, and disband semantic communities in a P2P environment; and ii) the
definition of a community-aware query propagation strategy for effective distri-
bution of resource requests on a semantic basis to enforce coordinated sharing
of distributed resources within a committed community. The Ph.D. thesis has
focused till now on the formation of autonomous and self-organizing emergent
semantic communities of peers. In particular, we have defined a semantic hand-
shake process based on ontologies and ontology matching techniques to handle
consensus negotiation and peer community formation. In this context, ontolo-
gies provide a semantically rich representation of the shared resources and enable
peers to describe their involvement in one or more concepts of interest. The role
of ontology matching techniques regards the semantic affinity evaluation be-
tween concepts provided by different peers in order to assess the level of match
between nodes with similar interests. A key feature of our ontology matching
techniques is related to their flexibility that makes them suitable for coping with
the inherent dynamism of open systems, such as P2P systems [5].
The research methodology that is being applied for the Ph.D. thesis is based
on the following main phases: i) literature review with the aim at providing a crit-
ical comparison of the state of the art solutions for managing peer communities
and semantic routing in P2P systems, ii) conceptual design where requirements
and foundational aspects related to the Ph.D. issues are formally addressed, iii)
experimentation with the aim at validating the thesis results by means of simu-
lation on a number of real test cases, and iv) prototype implementation where a
P2P prototype tool is developed according to the results and final considerations
of the Ph.D. thesis work.
2 Related work
Relevant research work with respect to the Ph.D. thesis regards community man-
agement and semantic query routing in peer-based systems.
Community management in peer-based systems. The idea of supporting
peer communities by means of a semantics-based approach is at an initial stage
of research, and few proposals have been appeared in the literature. In [4] a
social collaboration model is proposed as the reference basis to develop the P2P
Kex platform (Knowledge Exchange System) for supporting knowledge shar-
1154 CAiSE'06 Doctoral Consortium
ing in federations of peers. In this system, knowledge is organized according to
a XML-based syntax and a semantic matching algorithm is adopted to manage
the different meanings provided by single peers and federations. In [8], peer com-
munities are introduced as a generalization of peer groups to realize an efficient
search query propagation strategy in a populated P2P space. Peer communities
are formed on the basis of string-based interests that are used to determine the
communities in which a peer would participate. Each peer adopts an escalation
technique to advertise its interests and to take part to new communities. Trust
and reputation information are used in [1] to define clusters of peers (i.e., com-
munities) capable of providing relevant documents with respect to a given query.
The network clustering emerges gradually by point-to-point interaction among
the peers with the highest reputation. Each peer classifies the documents to share
and constitutes its local knowledge represented by means of a concept hierar-
chy. Syntactic and structural matching techniques are adopted to compare an
incoming query with peer local knowledge in order to identify possible common
features. In [3], communities are defined as groups within or across organizations
who share a common set of information, needs, or problems. In this approach,
peer interactions (i.e., queries) are exploited to discover the communities and
to populate the SWRC+COIN community ontology which describe the typical
structure and the key entities of communities as well as their relationships. The
community ontology can be queried in order to identify the most relevant peers
with respect to a given request.
Original contribution of the thesis. We observe that most of the presented
approaches recognize peer communities as a possible solution for improving sys-
tem effectiveness (e.g., query propagation). String- or XML-based formalisms
are used by peers to describe knowledge and interests. Syntax-based matching
functionalities are generally provided even if the adoption of semantic matching
techniques is becoming a key factor during the community discovery and forma-
tion phase. In this context, the main contribution of the thesis work is related
to the development of a semantic handshake process for peer community forma-
tion capable of combining ontological descriptions of peer interests with dynamic
ontology matching techniques. Such a solution provides semantic matchmaking
capabilities in community formation that allow to overcome the limitations of
exact matching techniques adopted in most approaches by contemporary ad-
dressing the dynamism and flexibility requirements of peer-based systems.
Semantic query routing in peer-based systems. Semantic query routing
strategies are required to improve the performance and the effectiveness of dis-
covery and search processes for resource sharing in P2P systems. In [10], the
REMINDIN’ multi-step query propagation protocol is described, to enforce se-
lected propagation of queries by observing which queries are successfully an-
swered by other peers, by storing these observations, and by subsequently using
this information for peer selection. A similar approach is presented in [12] where
the Intelligent Search Mechanism (ISM) is introduced to provide an efficient and
scalable solution for improving the information retrieval problem in P2P systems.
CAiSE'06 DC 1155
Each ISM peer is composed of four basic elements: i) the profiling structure that
is used to store the most recent replies of each known peer, ii) the query sim-
ilarity function that is used to identify the similarity between different search
queries, iii) the RelevanceRank algorithm which exploits the profiling structure
to select the peers that can provide relevant answers with respect to a given
query, and iv) the search mechanism that is used to send the query to the se-
lected peers. In most recent work, some initial ideas to consider query routing
as an application of peer communities have been appeared. In [11], the poten-
tial applications of communities are discussed and classified in endogenous and
exogenous applications. Referrals networks based on a sociological metaphor are
compared with bipartite communities based on link analysis in order to show the
benefits of a collaborative approach for improving local performance in locating
service providers. Agents (i.e., peers) adaptively select their neighbors and their
query recipients by exploiting sociability and expertise information computed on
previous interactions. The choices performed by the agents cause communities to
emerge. Furthermore, the notion of P2P Semantic Link Network is introduced
in [13] to emphasize the need of typed semantic links specifying semantic re-
lationships between peers in order to maintain information about nodes with
similar contents. Each peer defines its own XML Schema (source schema) de-
scribing the contents to share and adopts SOAP-based messages to communicate
with the other members of the network. Semantic links are exploited with cycle
analysis and functional dependency analysis in order to select the query recipi-
ents according to the types of the semantic links.
Original contribution of the thesis. Current P2P query propagation algo-
rithms are essentially based on statistical observations and exploit, in some cases,
a shared ontology, often just a taxonomy. The main contribution of the thesis
work is related to the definition of the community-aware query propagation strat-
egy to drive the selection of the best recipients for a given query by exploiting
ontology knowledge of the sending peer. One important goal of the proposed
approach is to address emergent semantics requirements, by extending current
techniques to work in multi-ontology contexts and thus releasing the constraint
of having an initial common shared ontology.
3 Preliminary results
In [8], the notion of peer community is introduced as a generalization of peer
group involving peers that are actively engaged in sharing, communicating and
promoting common interests. By extending such a notion, we define a semantic
community of peers as a set of nodes which show a common interest in a given
topic and are organized in a structured way (e.g., a tree).
Definition of semantic community. Formally, a semantic community SC is a
5-tuple of the form: SC = hCID, ICard, M embers, SP olicy, Statusi, where:
– CID is the unique Community Identifier that characterizes the community
SC.
1156 CAiSE'06 Doctoral Consortium
– ICard is the community Identity Card. The ICard represents a subject
category or topic area of interest and is defined as an ontology. The use
of an ontology-based ICard provides a semantically rich description of a
given topic area of interest and allows the characterization of the common
interpretation (i.e., perspective) featuring the community.
– M embers is the set of participants that joins SC and spontaneously agrees
with its ICard, since they have semantically relevant resources for the com-
munity.
– SP olicy ∈ (strict | sof t) defines the behavior that SC members have to
observe in terms of resource availability. The strict policy requires that in-
coming requests are processed by all community members in cooperation.
The sof t policy defines that each community member can autonomously
choose the set of incoming queries to evaluate.
– Status ∈ (potential | emerging | partially committed | committed | dis-
banded) represents the actual status of SC. During the consensus negotia-
tion process, the community passes through the potential, emerging, and
partially committed states. The committed and the disbanded states in-
dicate that the community is effective and no more active in the network,
respectively.
In our approach, we assume that each peer exposes to the system a peer on-
tology which provides a semantically rich representation of the resources that
the peer exposes to the network, in terms of concepts, properties, and semantic
relations [6]. Furthermore, each peer relies on the H-Match semantic match-
maker for matching ontologies in order to find which concepts match in different
ontologies and at which level [5].
In the following, we discuss some preliminary results regarding i) semantic
community formation based on the handshake techniques; and ii) the community-
aware query propagation.
3.1 Semantic community formation
As described in Figure 1, a semantic community of peers emerges when a node,
called community founder, invokes a semantic handshake process which is com-
posed of the following transitions:
1. ICard advertisement. The founder Pf defines a CID and an ICard describing
the topic area of interest of the emerging community, along with a set of com-
mitment constraints specifying the conditions required for the community
establishment (e.g., minimum number of member required, specific semantic
affinity constraints). Then, the founder composes an Invitation Message con-
taining the CID and the ICard created, as well as a TTL parameter defining
the maximum number of hops allowed for the invitation propagation. Then,
the invitation message is sent to all Pf neighbors in order to advertise the
new community.
CAiSE'06 DC 1157
Member
identification
ICard
Potential advertisement Emerging
Community Community
Request
approval
Community
commitment Partially
Committed
Committed
Community
Community
Fig. 1. The state transition diagram of the handshake algorithm
2. Member identification. Each invited peer Pi invokes the semantic match-
maker in order to compare the incoming ICard with its peer ontology. Pi is
relevant for the community if the semantic matchmaker identifies concepts in
the peer ontology with a high affinity with the ICard. In this case Pi replies
to Pf with an Interest Message reporting the portion of its peer ontology
related to the matching concepts found to be relevant for the community by
the semantic matchmaker. Independently from the matchmaker results and
if T T L ≥ 0, Pi forwards the invitation message to all its neighbors, except
for the peer from which the message has been received.
3. Request approval. Receiving the interest messages, the founder Pf has to
evaluate which peers are admitted in the community. For this reason, Pf
invokes its semantic matchmaker and compares each peer ontology portion
received by the interested peers with its knowledge (i.e., its peer ontology).
For each candidate peer, the goal of this comparison is to evaluate whether
the provided knowledge matches the knowledge of the founder, and then to
assess whether they share a common perspective of the community interests.
If the matchmaker returns high matching results, Pf admits the peer in the
community and sends an Approval Message to the admitted peer.
4. Community commitment. Once the Request approval phase is completed,
the founder verifies that the commitment constraints are satisfied. In this
case, a Commitment Message is sent to all the admitted peers and the se-
mantic community is effectively established. If the committed constraints are
not satisfied, the founder stops the community formation. In this case, the
admitted peers wait for the commitment message until a predefined timeout
expires and the community is considered as disbanded.
Appropriate techniques are also defined to address the main events that may
occur during the semantic community life-cycle, such as insertion and deletion of
participant, unexpected peer failure, and community disband. For further details
1158 CAiSE'06 Doctoral Consortium
regarding the semantic community formation and management, the reader can
refer to [7].
3.2 Community-aware query propagation
Committed communities are the reference for improving search and discovery
capabilities in P2P networks. When a searching peer Ps needs to submit a query
Q to the system, the communities of peers are exploited to select the query
recipients that can provide resources matching the target. To this end, Ps exploits
its joined communities in order to discover whether their ICards are related to the
query target. Ps invokes the H-Match semantic matchmaker and evaluates the
semantic affinity between the query Q and the ICard of each joined community.
On the basis of H-Match results, we distinguish the following cases:
– Ps is member of one or more communities related to the query Q. For each
community found to be relevant, Ps sends the query Q to its semantic neigh-
bors in the community. Each receiving node Pr forwards the query Q to its
community neighbors except for Ps , and invokes its semantic matchmaker to
compare the query Q against its peer ontology in order to evaluate whether
it can provide relevant knowledge to send back to Ps . The forwarding mech-
anism is iterated until the query Q reaches each community member.
– No semantic affinity exists between the query Q and the ICard of the com-
munities joined by Ps . Q is sent to all the peers known by Ps according to the
routing protocol of the underlying P2P infrastructure 1 . Each receiving peer
invokes the semantic matchmaker and compares the contents of Q with the
ICards it owns in order to renew the community-aware query propagation.
4 Ongoing and future work
In this abstract, we have presented the thesis work we are undergoing for seman-
tic community formation and management in P2P systems. Future work will be
devoted to finalize the development of the community formation process and
to assess the effectiveness of the proposed community-aware query propagation
techniques. In particular, two main issues will be considered:
Semantic handshake techniques. For what concern the semantic handshake
algorithm, we plan to implement such a semantic community aggregation pro-
tocol and to develop appropriate commitment policies for allowing a community
founder to specify the requirements to be satisfied by the potential member
peers for the establishment of an emerging community. Moreover, we intend to
refine the actual handshake process in order to share the responsibilities be-
tween the founder and the community members during the formation process.
In particular, we are working on the definition of advanced consensus negotiation
techniques in which the community ICard is the result of an active negotiation
1
In this case, a peer-based semantic routing protocol can be used to define query
recipients. Some initial results on this topic can be found in [6].
CAiSE'06 DC 1159
process where the founder and the interested peers interact and discuss changes
to the community ICard until an agreement among them is established.
Community-aware query propagation. By using simulation techniques, we
aim at comparing traditional P2P query propagation strategies with the com-
munity-aware propagation algorithm where queries are sent to the communities
with the higher chance to provide relevant results according to the matching
results. Further experiments will regard the comparison of the community-aware
query propagation with a basic semantic routing protocol we have developed
within the Helios framework for ontology knowledge sharing in P2P systems [6].
Finally, we are interested in developing popularity-driven community aggregation
techniques, where a peer founder can advertise a new community on the basis of
queries sniffed in the network. When a great number of queries in the network
is due to similar requests, a peer can propose to found a semantic community
regarding such a popular topic.
References
1. A. Agostini and G. Moro. Identification of Communities of Peers by Trust and
Reputation. In Proc. of the 11th Int. AIMSA Conference, Varna, Bulgaria, 2004.
2. S. Androutsellis-Theotokis and D. Spinellis. A survey of peer-to-peer content dis-
tribution technologies. ACM Computing Surveys, 36(4), 2004.
3. S. Bloehdorn, P. Haase, M. Hefke, Y. Sure, and C. Tempich. Intelligent Community
Lifecycle Support. In Proc. of the Int. I-KNOW Conference, Graz, Austria, 2005.
4. M. Bonifacio, P. Bouquet, G. Mameli, and M. Nori. Peer - Mediated Distributed
Knowledge Management. In Proc. of the Int. AMKM Symposium, Stanford, CA,
USA, 2003.
5. S. Castano, A. Ferrara, and S. Montanelli. Matching Ontologies in Open Networked
Systems: Techniques and Applications. Journal on Data Semantics, V, 2006.
6. S. Castano, A. Ferrara, and S. Montanelli. Web Semantics and Ontology, chapter
Dynamic Knowledge Discovery in Open, Distributed and Multi-Ontology Systems:
Techniques and Applications. Idea Group, 2006.
7. S. Castano and S. Montanelli. Semantic Self-Formation of Communities of Peers.
In Proc. of the ESWC Workshop on Ontologies in Peer-to-Peer Communities, Her-
aklion, Greece, 2005.
8. M. Khambatti, K. Dong Ryu, and P. Dasgupta. Peer-to-Peer Communities: For-
mation and Discovery. In Proc. of the PDCS Conference, Cambridge, USA, 2002.
9. P. Mika. Social Networks and the Semantic Web. In Proc. of the IEEE/WIC/ACM
Int. WI Conference, Beijing, China, 2004.
10. S. Staab, C. Tempich, and A. Wranik. REMINDIN’: Semantic Query Routing in
Peer-to-Peer Networks based on Social Metaphors. In Proc. of the 13th Int. WWW
Conference, New York, NY, USA, 2004.
11. P. Yolum and M.P. Singh. Dynamic Communities in Referral Networks. Web
Intelligence and Agent Systems, 1(2), 2003.
12. D. Zeinalipour-Yazti, V. Kalogeraki, and D. Gunopulos. Exploiting Locality for
Scalable Information Retrieval in Peer-to-Peer Networks. Information Systems,
30(4), 2005.
13. H. Zhuge, J. Liu, L. Feng, and C. He. Semantic-Based Query Routing and Het-
erogeneous Data Integration in Peer-to-Peer Semantic Link Networks. In Proc. of
the Int. Conference on Semantics of a Networked World, Paris, France, 2004.