=Paper=
{{Paper
|id=Vol-1312/ldop2014_paper5
|storemode=property
|title=Community-Driven Approach to Large-Scaled Ontology Development based on OAM Framework: a Case Study on Life Cycle Assessment
|pdfUrl=https://ceur-ws.org/Vol-1312/ldop2014_paper5.pdf
|volume=Vol-1312
|dblpUrl=https://dblp.org/rec/conf/jist/TakhomBISS14
}}
==Community-Driven Approach to Large-Scaled Ontology Development based on OAM Framework: a Case Study on Life Cycle Assessment==
https://ceur-ws.org/Vol-1312/ldop2014_paper5.pdf
Community-Driven Approach to Large-Scale Ontology
Development based on OAM Framework: a Case Study
on Life Cycle Assessment
Akkharawoot Takhom1, Prachya Boonkwan2, Mitsuru Ikeda3,
Boontawee Suntisrivaraporn1, and Thepchai Supnithi2
1 School of Information, Computer and Communication Technology
Sirindhorn International Institute of Technology (SIIT), Thammasat University, Thailand
akkharawoot.t@gmail.com, sun@siit.tu.ac.th
2 Language and Semantic Technology Laboratory
National Electronics and Computer Technology Center (NECTEC), Pathumthani, Thailand
{prachya.boonkwan, thepchai.supnithi}@nectec.or.th
3 School of Knowledge Science
Japan Advanced Institute of Science and Technology (JAIST), Nomi, Japan
ikeda@jaist.ac.jp
Abstract. This paper presents a community-driven development framework for
large-scale ontology that offers intra- and inter-community communication, vot-
ing and endorsement system, and version control. Our system design addresses
three lacks in the traditional ontology development: (1) constructive communi-
cation among the relevant stakeholders, (2) consensual endorsement and voting
system for concept and structural augmentation, and (3) bookkeeping and ver-
sion control. To cope with these shortcomings, we introduce three more tiers:
user management with technical profiles, knowledge augmentation via a web-
board, and community collaboration via the vote and endorsement system, on
top of the existing Ontology-based Application Management Framework
(OAM) [1]. When applying the design to the task of Life Cycle Assessment, we
discover additional needs to partition a large ontology to modularize the users’
responsibilities and to provide thread-based conversation for keeping track of
the topics.
Keywords: ontology development framework, community-driven approach,
large-scale ontology, knowledge management, version control, Life Cycle As-
sessment
1 Introduction
Life Cycle Assessment (LCA) is a kind of measures of environmental impact which is
used in environmental impact assessment (EIA). Offering opportunities to environ-
mental decontamination, it quantifies energy and materials that are consumed in the
process and released to the environment, and assesses the environmental impacts of
human activities and industries. The comprehensive knowledge of LCA is large-scale,
voluminous, and industry-specific. It has to be acquired from the published standard
documentations (ISO 14040 [2] , 14044 [3], and 14048 [4]), and numerous relevant
case studies, resulting in a very steep learning curve for the novice users. Develop-
ment and management of such knowledge are laborious and require a close collabora-
tion among experts. Therefore, knowledge transfer in the relevant stakeholder com-
munity becomes non-trivial in this task.
For ease of knowledge transfer, parts of LCA knowledge have been elicited in
terms of ontologies such as CASCADE [5], LCAO [6], Semantic Oil [7], and O-LCA
[8, 9], and transcribed into several representations, such as Semantic Web [10] encod-
ed as OWL [11], Predicate Logic, and Description Logic [12]. The notion of ‘ontolo-
gy’ promotes common understanding among the experts, makes the knowledge scala-
ble and reusable, and improves the interoperability between the experts and the users
and across relevant applications. Despite precise elicitation, the development of a
large-scale ontology is still an immense challenge because it requires effective
knowledge management when the entire community of experts are involved.
This issue can be alleviated by the notion of community-driven development ap-
proach. In this approach, it is assumed that every stakeholder is willing to participate
in the development and maintenance and contribute to the community with his exper-
tise. We focus on multidisciplinary knowledge integration and cross-checking among
domain experts and relevant stakeholders. A large-scale ontology is constructed and
maintained on a common platform on which the community of experts, knowledge
engineers, and users controls the development and maintenance processes, where each
decision making is done based on the community’s consensus. The platform provides
a means to communicate, transfer tacit knowledge, and discuss the changes of the
worked ontology with respect to immediate needs. Community-driven development
has been used in several ways, such as ontology matching [13, 14] and knowledge
curation [15]. There have been attempts to large-scale ontology development [16–20].
The community-driven approach is suitable for the development of LCA ontology
because of the following reasons. First, the domain experts specialize in their particu-
lar subfields of LCA. Since these fields sometimes share common knowledge, cross-
checking becomes necessary in a large-scale development project. Second, the devel-
opment of the LCA ontology is operated by a group of experts in parallel. In practice,
they usually branch (or fork) the current version of the ontology to work on their own.
This causes problems in updating when the finished ontologies are to be merged back
to the main ontology; thus, the need of version control. Third and last, some parts of
the ontology have to be cross-checked by specialists from other relevant fields; for
example, some parts of the ontology regarding earth and water can also be validated
by geophysicists, chemists, and environmentalists.
We propose the use of the Ontology-based Application Management Framework
(OAM) [1] as a community-driven development platform of the LCA Ontology. The
current framework offers the following facilities:
1. an ontology management tool that allows the community to get involved in
the evolution of the ontology,
2. a sandbox toolkit with application templates for a semantic search engine and
a recommender system, where no programming skills are required, and
3. APIs and web service deployment for practical application development.
However, it lacks collaborative features that enables constructive communication
among the stakeholders. The contribution of this paper is: we introduce to OAM the
notion of modified webboard, where any concept and structural augmentations are
proposed by experts and they have to be endorsed by the community. This allows the
system and the knowledge to grow along with the users’ expertise.
In this paper, we will show that the OAM Framework does not fully support com-
munity-driven development of a large-scale ontology where constructive communica-
tion is needed (section 2). Then we will show that our modified webboard enables
such communication oriented by concept and structural augmentations and consensual
endorsement (section 3). We will also explain how to apply this technique to the field
of LCA (section 4).
2 Problem Statement
OAM Framework [1] is a software platform that aims to simplify the development
and maintenance of a semantic web and an ontology as well as to automate the im-
plementation of a semantic search and a web service. The architecture of OAM is
illustrated in Fig. 1.
Ontology development via OAM entails three fundamental steps and three user
roles: domain experts, a knowledge engineer, and application developers. First, a
domain expert designs his own ontology according to the task of interest and export it
in the OWL format. On OAM, the knowledge engineer maps the ontology designed
by the experts to the database schema and the vocabulary and imports it into the sys-
tem. At this step, the knowledge engineer also maintains the current ontology accord-
ing to the experts’ requests via personal communication.
Fig. 1 System architecture of OAM Framework [1]
Second, the domain experts design recommendation rules for the current ontology
in terms of Prolog-like first-order logic. The knowledge engineer then transcribes
these rules into JENA Language via the Recommendation Rule Management Module.
Finally, the knowledge users implement their own knowledge-enhanced applications
with the Application Configuration Module. At this stage they can deploy a semantic
search engine and a web service using the ontology developed by the domain experts.
The key struggle in this paradigm is the ontology development entirely relies on
personal communication. This method is prone to the loss of communication; i.e. it is
very hard to keep track of conversations and consensus as time goes by. The history
of development evolution, or version control, plays a crucial role in community-based
development, especially for a large-scale ontology, in which a group of domain ex-
perts, knowledge engineers, and knowledge users are involved. These lacks necessi-
tate the use of a tractable communication means where conversations and consensus
are structurally organized for ease of bookkeeping, knowledge transcription, version-
ing, and deployment.
We are proposing an extension of the OAM Framework that incorporates the no-
tion of thread-based webboard, version control, and status notification to solve the
aforementioned problems. These features allow the community with the three user
roles to co-create a large-scale ontology and maintain it by means of community en-
dorsement. By doing so, the system and the knowledge grow along with the users’
expertise.
3 System Design
In this paper, we design a community-driven ontology-based application management
framework (CD-OAM). We extend the canonical OAM Framework (shown as Data
Tier and Application Tier, all in dashed borders) with three more tiers: Collaboration
Tier, Knowledge Tier, and User Tier (all in solid borders). A system overview of CD-
OAM is illustrated in Fig. 2.
Fig. 2 System Overview
First, the User Tier facilitates the admin to manage domain expert members, in-
cluding adding, deleting, and assigning to some of the communities. The users are
required to create a profile annotated with their expertises so as to classify them with
respect to their interests, technical backgrounds, and objectives. Our user policy is
merit-based [14]: the more a user participates in votings or change proposals, the
more merits he gets.
Next, The Collaboration Tier, the heart of this framework, encourages the commu-
nity-driven development of a large-scale ontology. It contains three modules: com-
munity management, vote and endorsement system, and notification system. The
Community Management module allows the community’s administrator to manage
their members and their user roles. The Vote and Endorsement System prepares a
platform for voting and endorsement by the community when changes in the ontology
take place. Finally, the Notification System disseminates the voting results and
changes to the community.
Finally, the Knowledge Tier is composed of four modules that facilitates both in-
tra- and inter-community communication in large-scale ontology development and
keep track of the evolution. We assume that a large-scale ontology can be separated
into parts and distributed to all involved communities. The Intra-community Ontology
Management Module [15] allows the community to manipulate the part of the ontolo-
gy the community is responsible for. The Inter-community Knowledge Augmentation
Module facilitates the integration of each part of the ontology. Should modification is
necessary, this can be voted and endorsed by all relevant communities. All proposed
changes must be endorsed by the community. The Webboard Module organizes a vote
for intra- and inter-community changes in the ontology and approves the consensus.
For example, an expert may propose to augment some concepts or the hierarchy in the
Webboard. If changes are endorsed and committed, the history is kept in the Version
Control Module.
4 Case Study: LCA
Ontology development for LCA is a very challenging task. Standardized by the ISO
into ISO 14040, 14044, and 14048, LCA is a family of best-practice procedures for
information sharing and guidelines for environmental impact evaluation. It minimizes
operations in the organization which affect the environment, to comply with laws,
regulations, and other environmentally oriented requirements and continual improve-
ment. Figure 3 shows the ISO guideline for LCA that consists of four main phases:
(1) setting the goal and scopes (2) listing up life cycle inventory from a given supply
chain (3) assessing the life cycle impacts and (4) data interpretation. The ontology for
LCA is usually very large because there are numerous concepts and a complicated
hierarchy and relations between them in the product’s life cycle.
Fig. 3 ISO standards guideline of LCA
Due to its sheer scale, there are a number of stakeholders working side by side on
the ontology. We classify them into five categories with respect to their objectives:
1. Domain experts: experts and specialists on LCA
2. Domain beginners: beginners who just started in the field and are learning
the knowledge
3. Knowledge users: direct users of the knowledge via a recommender sys-
tem
4. System developers: software developers who make use of the knowledge
5. Knowledge engineers
Complying with the community-based approach, these stakeholders have different
roles in developing the LCA ontology and they all need a means to communicate to
each other. Any requests of change in the ontology have to be endorsed by the entire
community. Moreover, version control becomes necessary as the ontology grows by
the community. Our system design fit all these requirements for the community-based
LCA Ontology Development.
Our system design has to be specialized as per the roles and interests of the users.
We envisage the following modifications in the following tiers when applied to the
field of LCA.
User Tier: Each user should specify his objectives, role, interest, and tech-
nical backgrounds regarding LCA. These users are then assigned to a differ-
ent part of the ontology according to their request or technical backgrounds.
Collaboration Tier: To modularize the user responsibilities, the community
are able to partition (and repartition) the ontology via voting and endorse-
ment.
Knowledge Tier: There are at least two sources of knowledge in LCA: ISO
standards and field data collection. Knowledge integration becomes a non-
trivial issue because domain experts may have different perspectives on the
received data. The webboard module has to offer a place for open discussion
before leading to voting and endorsement. To keep track of the decision, a
thread-based conversation is best suitable for this scenario.
5 Conclusion and Future Work
We have presented a design for a community-driven development framework for
large-scale ontology, where all relevant stakeholders can participate in the evolution
of the ontology. Three additional layers: user management, knowledge augmentation,
and community collaboration, respectively, are put on top of the OAM Framework to
facilitate communications among the stakeholders, voting and endorsement, and ver-
sion control. We also find that, in practice, ontology partitioning and thread-based
conversation are also needed for teamwork as found in Life Cycle Assessment.
Our future work remains as follows. First, we will implement the system closely
following this design and conforming to User Experience (UX). Second, we will in-
corporate the notions of conflict detection and concept similarity into the knowledge
augmentation module. This will help knowledge engineers and the community prede-
termine holistic conflicts in the ontology before committing any changes. Third and
finally, we will develop a data modeling module [21] which automatically maps input
databases with a different data scheme to the ontology.
Acknowledgment
This research is financially supported by Japan Advanced Institute of Science and
Technology (JAIST), Sirindhorn International Institute of Techcnology (SIIT), and
Thammasat University (TU). Materials and information provided by National Metal
and Materials Technology Center (MTEC, Life Cycle Assessment Laboratory, Thai-
land). The authors are grateful to Marut Buranarach for fruitful technical discussions.
References
1. Buranarach, M., Thein, Y., Supnithi, T.: A Community-Driven Approach to
Development of an Ontology-Based Application Management Framework. In:
Takeda, H., Qu, Y., Mizoguchi, R., and Kitamura, Y. (eds.) Semantic
Technology. pp. 306–312. Springer Berlin Heidelberg (2013).
2. ISO 14040:2006 - Environmental management -- Life cycle assessment --
Principles and framework. ISO, Geneva, Switzerland (2006).
3. ISO 14044:2006 - Environmental management -- Life cycle assessment --
Requirements and Guidelines, (2006).
4. ISO/TS 14048:2002 - Environmental management -- Life cycle assessment --
Data documentation format, (2002).
5. Cappellaro, F., Masoni, P., Moreno, A., Scalbi, S.: CASCADE. In: Werner
Pillmann, K.T. (ed.) The 16th Internationale Conference: informatics for
environment protection. pp. 490–493. IGU/ISEP (2002).
6. Brascher, M., Monteiro, F., Silva, A.: Life cycle assessment ontology. the 8th
Conference of the International Society for Knowledge Organization. pp.
169–177. Congreso ISKO-España, España, Spain (2007).
7. Bertin, B., Scuturici, V., Risler, E., Pinon, J.: A semantic approach to life
cycle assessment applied on energy environmental impact data management.
Proceedings of the 2012 Joint EDBT/ICDT Workshops. pp. 87–94 (2012).
8. Takhom, A., Suntisrivaraporn, B., Supnithi, T., Theeramunkong, T., Manabu,
O.: Ontology-enhanced life cycle assessment: toward formalizing the standard
guidelines. The International Conference on Information and Communication
Technology for Embedded Systems (ICICTES 2013). , Samutsongkhram,
Thailand, (2013).
9. Takhom, A., Suntisrivaraporn, B., Supnithi, T.: Ontology-enhanced life cycle
assessment: a case study of application in oil refinery. The Second Asian
Conference on Information Systems, (ACIS 2013) (2013).
10. Horrocks, I.: Ontologies and the semantic web. Commun. ACM. 51, 58–67
(2008).
11. Schreiber, G., Dean, M.: OWL Web Ontology Language Reference. (2004).
12. Baader, F., Horrocks, I., Sattler, U.: Description Logics. Stud. Health
Technol. Inform. 101, 137–41 (2004).
13. Zhdanova, A. V: Towards Community-Driven Ontology Matching. Computer
(Long. Beach. Calif). 1–2 (2005).
14. Zhdanova, A. V, Shvaiko, P.: Community-driven Ontology Matching.
Proceedings of the 3rd European Conference on The Semantic Web: Research
and Applications. pp. 34–49. Springer-Verlag, Berlin, Heidelberg (2006).
15. Groza, T., Tudorache, T., Dumontier, M.: Commentary: State of the Art and
Open Challenges in Community-driven Knowledge Curation. J. Biomed.
Informatics. 46, 1–4 (2013).
16. Gendarmi, D., Lanubile, F.: Community-Driven Ontology Evolution Based
on Folksonomies. Move to Meaningful Internet Syst. 2006 OTM 2006 Work.
181–188 (2006).
17. Hepp, M., Bachlechner, D., Siorpaes, K.: OntoWiki: Community-driven
Ontology Engineering and Ontology Usage Based on Wikis. Proceedings of
the 2006 International Symposium on Wikis. pp. 143–144. ACM, New York,
NY, USA (2006).
18. Siorpaes, K.: Lightweight community-driven ontology evolution. Lecture
Notes in Computer Science (including subseries Lecture Notes in Artificial
Intelligence and Lecture Notes in Bioinformatics). pp. 951–955 (2007).
19. Aseeri, A., Wongthongtham, P.: Community-driven ontology evolution based
on lightweight social networking in oil and gas domain. IEEE International
Conference on Digital Ecosystems and Technologies. pp. 197–202 (2011).
20. Maleewong, K., Anutariya, C., Wuwongse, V.: A Semantic Argumentation
Approach to Collaborative Ontology Engineering. Proceedings of the 11th
International Conference on Information Integration and Web-based
Applications and Services. pp. 56–63. ACM, New York, NY, USA (2009).
21. Bertin, B., Scuturici, V.-M., Pinon, J.-M., Emmanuel Risler: A semantic
approach to life cycle assessment applied on energy environmental impact
data management. Proceedings of the 2012 Joint EDBT/ICDT Workshops.
pp. 87–94. ACM, New York, NY, USA (2012).