Content uploaded by B. T. Sayed
Author content
All content in this area was uploaded by B. T. Sayed on Sep 21, 2018
Content may be subject to copyright.
Towards Designing the Info-structure for Developing Socio-technical
Knowledge Management System (KMS)
S.M.F.D Syed Mustapha
Department of Computer Science
Dhofar University
Sulatanate of Oman
smfdsm@yahoo.com
B. T. Sayed
Department of Computer Science
Dhofar University
Sulatanate of Oman
btsm999@yahoo.com
Abstract
It can be acknowledged that the notion about the importance of Knowledge
Management (KM) has been widely accepted and many researchers have
devoted into making it a successful story. E ve n tho ug h it is a broad and
multidisciplinary subject in nature, researchers have been working on a well-
focused area of the entire scope in KM . Many of the researchers have made a
successful claim about their research outputs. Nevertheless, it is also being
commented that the KM when fully implemented in a real world environment
does not show an overall success as expected. The diversities of research
attempt by different research groups have made them to ignore many aspects
which are beyond their interest. This paper reviews cradle-to-grave aspects and
categories of knowledge management research that ranges from technical to
social issues. Many aspects of t he research that have been investigated, the
research outputs are sufficiently matured to be used. In order to build a
knowledge management system (KMS) that balances all aspects of its
components, a socio-technical KMS is proposed in this paper. The components
of KM are the people, the technology and the resource. We classify the research
work that addresses each of these components as well as the overlaps of any of
the two. Eventually, the socio-technical system that lies at the nexus of the three
is shown and illustrated.
Introduction
Knowledge management is still a growing field of research and has attracted
researchers from multidisciplinary areas. The classical notion about knowledge
management as a systematic repository system for storing and retrieving knowledge
sources has encroached to a broader definition, functionalities and application areas.
The research findings and outcomes from these areas can be deployed to enhance the
sophistication of the knowledge management system. Various people who are involved
in part of knowledge management either as system developers or users are the
managers, scientists, computer technologists, medical doctors and psychologists. Their
interests and directions may seem to differ by the nature of the needs and problems that
they are looking at. They may overvalue certain aspects of KMS component and
downplay the others and therefore this consequently can create imbalances of factors.
There are three main factors to consider in the entire process that make so-called the
Knowledge Management System (KMS) which are the people, the technology and the
resource.
This paper reviews the research areas that focus on each of these factors as well
as looking at the overlaps of these factors as shown in Figure 1. The people are the
main KMS subject where knowledge is destined, processed, transformed, reproduced
and regenerated. Besides, the network of the people that forms a community provides a
larger composition of knowledge community in the entire KMS component. The
technology is the techniques, models, algorithms, architectures and devices that creates,
facilitates, supports and operates the knowledge within the KMS. The resource defined
in KMS has broader sense that it includes any objects that intermediates between the
knowledge providers to the knowledge recipients. That means, it does not only include
the ones that are kept in printed media or electronic format such as the newspaper,
magazines, articles, brochures, manuals, operating procedures, web pages, videos, e-
mails, graphical animations but also the artifacts that are part of learning environment.
For example, the animals in the zoo can be the resource to the zookeepers in learning
about animal behaviors by observing them. The colleagues and friends are the common
resource of someone who frequently refers to in getting their job done.
There are also researches that are double flank shown to appear in both sides such as
T&P (Technology and People), R&P (Resource and People) and T&R (Technology and
Resource). In T&P, the technology is designed by considering highly on human factors
and needs that without them the technology remains useless. In R&P, the people
interact directly with the resource without using the technology as the interface. The
resource actuates the thinking process, the ideology, the principles, the decision styles,
the social inclination, the cultural adaptation and personality of the people. For example,
a news break about economic crisis in Asia turns away potential investors quickly when
People
(Community)
Technology
Resource
R & P
T & P
T & R
R & P &
T
Figure 1 Knowledge Management System components
they read about it. In T&R, the technology is developed with an association with the
resource in making the resource easily to search, to compile, to summarize, to represent,
to organize, to store (considering the semantic and syntactic approaches) and to
manipulate for regenerating a new form of resource. At the nexus of all KMS
components are the R&P&T which consider all aspects and balance these factors in the
process of developing a socio-technical KMS.
In the subsequent sections we discuss the existing research that focus the
aspects discuss above and finally conclude with an illustration of socio-technical KMS
that we are working on in our Intelligent Conversational System (ICC).
Technology
Most researchers that are working on technology-focused research will eventually apply
the technologies in the real world problem. It is therefore difficult to disassociate them
with people or resource in this context with a clear border. Nevertheless, for the purpose
for categorization in this discussion, we consider them to be a purely technological when
they are problem-independent, versatile in usage and they emerge prior to the existence
of the problems. These technologies are highlighted in this paper without elaborative
details as their descriptions are easily available elsewhere.
Latent semantic indexing is an applied method to the singular value
decomposition (SVD) which is the fundamental eigenvector decomposition technique for
matrices [1]. Matrices are used to represent the structure of complex relationship s i n
which in the case of document retrieval is the document to the terms. It resembles factor
analysis technique that is also largely used by statisticians. The singular value
decomposition is used to reduce the dimensions of vector space. The output of the SVD
describes the location of the terms geometrically in k-dimensional spaces.
The techniques in machine learning that are frequently used are the supervised
and unsupervised learning [2,3]. Both techniques are unique in their strengths and
weaknesses depending on how they are used. Supervised learning requires predefined
input pattern for the neural network to learn such that it is more suitable for classification
task. When the desired pattern for the set of input is hardly known, the unsupervised
learning is more appropriate for categorization task. Someone working on identifying
different types of randomly chosen topic reports will use unsupervised learning in
comparison to a medical doctor who is interested to select reports of medical relevance.
Natural language processing (NLP) has been resorted in the information
extraction task [4]. NLP is more favorable option as a technique to identify a specific
grammatical meaning of a term in comparison to LSI that recognizes the document by
the general structure of the terms. NLP builds a tree-like structure of a sentence and has
a parsing technique that produces the grammatical meaning of words by tokenizing and
tagging them. In furtherance to getting the real world meaning of a word such as name
of location and people, types of diseases and category of tasks and actions, domain-
specific knowledge is incorporated [5]. For example, when Kuala Lumpur is identified as
noun by the NLP tagger, this name is searched in the domain knowledge to be identified
as the name of a city of a country.
Resource
Resource is the learning object that is used as a media in transmitting information as
well as a source for learning. The generality of its definition broadens the choices o f
objects that are qualified to be considered as the resources. Having said that, the
examples of resource are not only limited to newspaper, magazines, articles but also
other objects that intermediate between the processes of learning. In many occasions,
colleagues are the main form of learning object or at least the highest priority choice of
reference when immediate clarification is needed. In the professional fields, the design
plan of the building and the building structure photos are the learning objects for the
architects; the animals and their habitats are the learning objects for the zookeepers; the
meeting minute, rules and procedures and report are the learning objects for the top
management in making decisions. With the advent of multimedia technologies, the
resource can be represented online and stored in the KMS repository.
People
Researchers in KM have now agreed that people are the essential entity that makes the
knowledge management works regardless how simple or sophisticated is the technology.
With simple technology such as e-mail, humans have benefited a lot when they
extensively use it. In KMS, there are three main things that are relevant about the people
which are the social communication, social learning and social knowledge. Social values
must exist to form the community network of an organization as a prerequisite to KMS.
The values are measured by detecting the quality of interactions, sharing, motivations,
commitment, ideas and political settings. Social communication reflects the common
language, buzzword, cliché, jokes and understanding which can be foreign to the
apprentices in the organization or outsiders. For example, the social communication of a
person with his colleague who he has been working together for a long time will differ to
a client whom he just knew. A strong social communication creates smooth avenues to
interactions, sharing and commitment. Social learning emerges when there is a strong
social communication. People learn how to complete their tasks better by quickly
referring to the colleagues and also counterparts. For example, the ideas about vision,
goals and objectives of a company and the political settings are easily learn through
socialization. Social knowledge is about knowing the right people to contact, the
common practice of the organization (may not be the best practice), the organization’s
culture and beliefs that can only be known from the people.
Community of Practice (CoP) has been adopted widely as specifications to
identify the existence of learning community in an organization [7]. The CoP values are
essential as the foundation to the cohesiveness, oneness, convergence in thinking and
common in practice of the society.
Technology and Resource
A group in Carnegie-Mellon is working on Topic detection and tracking system
[7,8]. It processes a series of stories in order to detect distinct events from the stories.
The research aims to be able to extract significant events, to determine if there is any
recent update from the events and to summarize the events for the user. Two techniques
which are the supervised and unsupervised learning are used. The supervised learning
is used for a sample story that is identified by the user. The unsupervised learning allows
the system to form the classified group of events automatically.
Another group of similar interest from Universität Dortmund uses support vector
machine to categorize text of different types into a predefined category [9]. The text
consists of stories of different topic of some categories. The approach is purely
automated without using the human intervention along the process.
In different effort, web services are used as the technology to represent the
semantic content of the resource using web language technology [10]. DAML+OIL and
OWL are web ontology languages that are used to represent the logic reasoning where
XML is the basic syntax. The attempt aims to extend the web capability in supporting the
intelligent system running on the internet at the client side rather than the server itself.
Video is also one of the learning objects as it captures the reality of events in a
recorded format. Therefore, it has become a preferential choice for the university to use
in distance learning, online meetings during teleconference and digital archives. The
techniques for storage, compression, indexing and retrieval have intruded into the
lifecycle of video as it is massively used and that requires systematic approaches.
Research in content-based video retrieval is concerning about automating the video
content recognition by analyzing the spatial scene and temporal dimension [11]. By
having the ability to segment the video content, the video slices can be indexed and
stored for easy retrieval.
Document understanding is another interesting effort in integrating technology
and resource. The research group in University of Amsterdam had attempted to
automate the process of recognizing the reading order of a document [12]. It adopts
techniques from computer vision, artificial intelligence and natural language processing.
The system uses generic approach but it is still able to handle broad class of documents.
Technology and people
There are many areas that we can consider the research to fall under this
category. It ranges from developing the system to support communication, delivering
personalized information, to facilitate group learning and to model the user’s acceptance
to technology. The researchers working in this area have a common argument that the
sophistication of technology will not attain the underlying objective in KM without the
commitment of the people [13].
Some obvious examples are the e-mails, chatting rooms and bulletin boards
where contribution from the people is required in order to claim the system to be
successful.
Collaborative knowledge building emphasizes on human’s collaborative effort in
knowledge shaping [14, 15]. Gerald Stahl is looking at this approach where his system
requires a community of people involve in contributing ideas through social
constructivism. Collaborative learning has similar approach but focusing on specific
group of participants like students.
Thomas Erickson [16] builds Babble system to facilitate and catalyze social
interactions of a community. The community involvement is shown visually in terms of
active participants by representing them symbolically on the coordinate space. Less
active participants can be seen to be away from the cluster of symbols and this
subsequently encourages them to response accordingly.
Yogesh Malhorta and others measure empirically the user’s opinion about
knowledge management using Technology Acceptance Model (TAM) [17]. It models the
relationship of the psychological variables such as the perception, the attitude and the
behavior of the computer usage of the users.
Resource and people
The research in resource has been a long-standing research issues even before the
advent of its technology. The scope is broad that it encompasses the design issues, the
pedagogical issues, the delivery methods, the learning strategy and most importantly the
effectiveness in grasping reader’s understanding. In relation to the earlier discussion on
Technology and Resource, the resource is now enlivened with additional technological
features. Consequently, it has been taken into granted in enhancing the teaching,
learning and assessment methodology. As a result, the new form of resource has now
created new relationships with the people who use it. Computer-aided learning allows
interactions between the resource and the learner that the latter can choose the learning
path in more flexible way and will receive responses when question is posted. Intelligent
tutoring system compartmentalizes the structure and organization of the resource such
that the communication, the knowledge representation, the pedagogical and teaching
aspects can be modeled to suit the content of the resource. Multimedia technologies
enrich the interaction capabilities with animations, sounds, hyperlinks, hypermedia,
images and video streaming.
A new direction of research in this area is looking at the social aspect of the
document which in our context is referred to the resource. This idea was highlighted by
John Seely Brown and others as describe in [18]. The document is argued to be able to
create political linkage among readers through reading the same material even though
they have not physically met. The content of the document is repetitively negotiated with
different meaning by the readers every time it is consulted.
Socio-technical KMS through Intelligent Conversational Channel
The term socio-technical has been used elsewhere to indicate the symbiotic of
the technology features as well as the people [19, 20]. Intelligent Conversational
Channel is a knowledge sharing tool that is built that we claim to balance the three main
components of KMS, namely the people, the resource and the technology. Before we
describe further about the system, the following questions are posted as the arguments
considering mainly the technical issues:
1. How accurate is the retrieval system should be in order to ensure the
satisfiable retrieved results?
2. What is the ontological framework that can be used in order to ensure it
represents well the variability of human’s knowledge?
3. Will high computational cost that have incurred in order to model
completeness of the algorithm worthwhile when what is needed and used
are still uncertain?
4. Can we build a full automated system without the need to pay the cost for
the cold start?
These questions are considered the major ones and related to the previous work in the
discussion given earlier even though there are many more. The attempt to retrieve
documents is purely based on mathematical constructs such as recall and precision in
which the number of terms occur in the documents are based on. This approach works
well if the selected documents are the ones preferred by the users but the entire setting
of the process do not require their involvement. For example, the topic detection and
tracking system can successfully locate the event boundary after processing 16,000
stories but users practically can not wait for such computational cost and time and the
events their interest towards event topic quickly change from time to time. Building
ontology framework requires great effort when new ontological language on the web is
still an ongoing process. Research in the large-scale knowledge based system had
shown it is a decade of human research effort and yet the application result is still at
minimum usage. Research in community of practice emphasizes that people acquire and
need the knowledge at work and with colleagues more than during the formal training
where such knowledge is called social knowledge. Research has found the importance
of getting users to involve in solving the cold start problems when initiating the right
samples to use before processing and also using the users for feedback after the
retrieval [21, 22].
Intelligent Conversational Channel has been fully described in previous
publication in terms of the architecture and implementation [23, 24]. In this paper, the
system is described based on the following definitions that reflect the characteristics of
socio-technical-based system.
a. Resource: Supports multi-facetted resources and knowledge resource
enrichment.
b. Technology: Facilitates community design on classification, categorization,
knowledge building and indexing, and supports community and social
network analysis.
c. People: Builds social cohesion among community in the organization and
catalyzes knowledge community through resource sharing and
community involvement.
Resource:
Recompilation
and revisit
Resource:
Gathering
Resource:
Sharing
Resource:
Analys is
Personal resource
gathering
Magazines,
Web page
Books, article
papers
Personal knowledge
background
Experience,
training and skills
Observation,
readings and
conversation
Rebuilding of
personal
resource
Rebuilding of
personal
knowledge
background
Facts accrual and
affiliation
Selection of
facts
Mu ltip le
resource
analysis
Share without
bias
Share without
unnecessary
retention of
knowledge
Ethical and
professionalism
Figure 2. Resource activities for socio-technical system
ICC
Figure 2 shows the community is actively involved in the resource activities in
terms of gathering, analysis, sharing and, recompilation and revisit. The
community designs their personal resource collections based on their own
personal knowledge background and interest. Besides physical resource
collections, they earn knowledge through experience, training and skills. The y
analyze multiple resources through reading the relevant parts of knowledge
inquisition, making selection of relevant materials and accruing the facts affiliated
to the subject matter. The knowledge is shared then through Intelligent
Conversational Channel (ICC) in which the system adopts certain policies to
ensure etiquettes and professionalisms on t he internet is maintained to prevent
flaming and internet abuse. The resources are shared in multi-facetted form such
as articles, videos, web documents and graphical pictures as well as personal
opinion and experience through storytelling. The resources are built
collaboratively by the community and consequently these resources are
recompiled and revisited by the community to rebuild their own personal resource
repository.
People:
Social Communication
Social Learning
Symbolic Convergence
Pe op le:
Community of
Practice
Peop le:
Virtual Community to
support actual
community
Figure 3. People activities in Socio-technical system
ICC
It is known that internet-based communication channels involve people to
participate actively in the dialogue session. In the socio-technical system, active
participation is not only the primary requirement but also the existence of social
values is greatly emphasized. Community of practice theory [5] has been
adopted widely in many attempts in developing computer-mediated
communication system [23,24]. In the theory, the societal cohesiveness plays
important role in creating collegial learning. It is claimed that apprentice of an
organization learns how to perform duty with the colleagues most of the time
through the social communication. We extend this idea to include social learning
and symbolic convergence as the basis to achieve a successful knowledge
sharing environment (extensively explain in [25]). With regard to the limitation of
time and space in engaging the community to participate in the synchronous
online session, agent technology is embedded in the ICC system [26]. The
agents represent the social characteristic of the actual communities when they
are absent during the dialogue session with another communities.
Figure 4 shows the symbiotic between technology and community effort
that simplify the computational complexities and reduce cost. The technology as
shown requires searching of resources using available search engines. Using the
community’s intelligence, the resources are classified and categorized in the
evolutionary manner. The current effort to automate this task is useful when there
are volumes of resources to be processed in a short time. However, the capacity
of reading by individuals is much lower than the amount of retrieved resources.
The community posted short messages to the ICC system together with the
learning artifacts which are associated to it. The short messages that consist of
10 to 50 words are stemmed to remove the functional words a nd the y are stored
as the indices which can be referred to for retrieval purpose. Social Network
Analyzer keeps statistical record on the participation profile of the community in
terms of their social pattern and communication behavior. ICC provides several
channels for different topics of discussion such that the automation of topic
detection is unnecessary.
Technology:
Searching
Technology:
Building knowledge
repository and
indexing
Technology:
Clas sificatio n &
Categorization
Technology:
Social network
Analys is
Googles and
Yahoo search
engine
Mu ltip le u se r
effort
Mu ltip le
documents
selected by a
user
Word stemming
Contextual
indexing
Pa rtic ipatio n
profile
Quality of
participation
Technology:
Social network
Analys is
Channel of
discussion by
topic
ICC
Artifacts are
associated with
posted message
Figure 4 Technology involves in Socio-
technical system
Conclusion
This paper discusses the current research effort in the making of successful
knowledge management system. Each research focuses on different component
of knowledge management aspects. Despite we agree that these efforts are
headway to the sophistication of KMS, the balance consideration between the
needs of the people who use it, the minimum technology required to enable the
KM activities and the enrichment of multiple form resources is argued in this
paper. The research efforts are classified into six categories where three of them
are double flank of KM components. We propose the socio-technical KMS that
centers between the people, resource and technology and argue that these
concepts do not need highly sophisticated KMS at the same time the targets of
KM are achievable.
References
1. Husbands, P., Simon, H., and Ding, C. On the use of singular value decomposition
for text retrieval. In M. Berry, editor, Proc. of SIAM Comp. Info. Retrieval Workshop,
October 2000.
2. Dingli, A., Ciravegna, F., Guthrie, D., and Wilks, Y. Mining Web Sites Using
Unsupervised Adaptive Information Extraction. In Proceedings 10th Conference of
the European Chapter of the Association for Computational Linguistics, Budapest,
Hungary, 2003.
3. Mostafa, J and Lam, W. Automatic classification using supervised learning in a
medical document filtering application. Information Processing and Management, 36,
pp 415-444, 2000.
4. Kurland, O. and Leen, L. Corpus structure, language models, and ad hoc
information retrieval, Proceedings of SIGIR, pp 194-201, 2004.
5. Wenger, E. Communities of Practice. Cambridge, Eng: Cambridge University Press,
1998.
6. Yang, Y., Carbonell, J., Brown, R., Pierce, T. and Archibald, B.T., and L., Xin.
Learning approaches for detecting and tracking news events. IEEE Intelligent
Systems, 14(4):32--43, 1999.
7. Allan, J., Carbonell, J., Doddington, G., Yamron, J., and Yang, Y. Topic detection
and tracking pilot study: Final report. In Proceedings of the DARPA Broadcast News
Transcription and Understanding Workshop, 1998.
8. Joachims, T. Text categorization with support vector machines: Learning with many
relevant features. Proceedings of ECML-98, 10th European Conference on Machine
Learning, 1398, Springer Verlag, Heidelberg, DE, pp 37-142, 1998 .
9. McIlraith, S and Martin, D. L. Bringing Semantics to Web Services. IEEE Intelligent
Systems, 18(1):90--93, January/February, 2003.
10. Müller, H., Müller, W., Squire, D. M., Marchand-Maillet, S., and Pun, T. Performance
Evaluation in Content-Based Image Retrieval: Overview and Proposals, in Pattern
Recognition Letters (Special Issue on Image and Video Indexing), H. Bunke and X.
Jiang Eds., 5, 22, pp 593-601, 2001,
11. Aiello, M., Monz, C., Todoran, L., Worring, M. Document understanding for a broad
class of documents. International Journal on Document Analysis and Recognition, 5,
pp 1 – 16, 2002.
12. Malhotra, Y. and Galletta, D. F. Building system that the user wants to use.
Communications of the ACM. Volume 47, Number 12 (2004), pp 88-94
13. Stahl, G. Collaborative Information Environments to Support Knowledge
Construction by Communities..AI & Society,14: 71-97, 2000.
14. Erickson, T. and Kellogg, W. A. Knowledge Communities: Online Environments for
Supporting Knowledge Management and its Social Context. Sharing Expertise:
Beyond Knowledge Management (eds. Ackerman, Mark, Volkmar Pipek, and Volker
Wulf). Cambridge, MA, MIT Press, 2003, pp. 299-326
15. Malhorta, Y., and Galletta, D. F. Extending the Technology Acceptance Model to
Account for Social Influence: Theoretical Bases and Empirical Validation. Thirty-
Second Annual Hawaii International Conference on System Sciences-Volume 1,
Maui, Hawaii, 1999
16. Brown, J.S., and Duguid, P. The Social Life of Information Harvard Business School
Press, 2000, ISBN 0-85784-762-5.
17. Van House, N. A., Bishop, A. P., and Buttenfield, B. Introduction: Digital Libraries as
Socio-technical Systems. Ann P. Bishop, Barbara P. Buttenfield, and Nancy A. Van
House, eds, Digital Library Use: Social Practice in Design and Evaluation, MIT P res s ,
Fall, 2003
18. Douglas, K and Laurence, B. Applying social-technical approach for COTS selection.
Proceedings of 4th UKAIS Conference, University of York, Mc Graw Hill, April 1999.
19. Middleton, S., Alani, H., and De Roure, D.C. Exploiting Synergy between Ontologies
and Recommender System. In Proceedings of Semantic Web Workshop 2002 At the
Eleventh International World Wide Web Conference Hawaii, USA.
20. Rocchio, J.J. Relevance feedback in information retrieval, in the SMART Retrieval
System – Experiments in Automatic Document Processing, Englewood Cliffs, NJ,
Prentice Hall, pp 313-323, 1971.
21. S.M.F.D Syed Mustapha. Intelligent Conversational Channel for Learning Social
Knowledge among Communities. 8th International Conference on Knowledge-based
Intelligent Information & Engineering System, KES 04, New Zealand, Vol I, pp343-
349, 2004
22. S.M.F.D Syed Mustapha. Agent-mediated Community Channel for Social Knowledge
Building in Educational Environment. 5th Int. Conf. on Information Technology Based
Higher Education and Training: ITHET ’04, 31st CD proceeding, Istanbul, Turkey,
Ma y – 2 June, 2004.
23. Kimble, C., Hildreth, P., and Wright, P. Community of Practice: Going Virtual. In
Knowledge Management and Business Model Innovation. London: Idea Group
Publishing. pp 220-234. Available from : http://www-
users.cs.york.ac.uk/~kimble/research/13kimble.pdf.
24. George, J.F. Learning in Context: Extensively Computerized Work Groups as
Communities-of-Practice. Accounting, Management and Information Technology.
5(3/4), 185-202, 1995.
25. S.M.F.D Syed Mustapha and B.T. Sayed. Knowledge sharing pyramid:
Communication, Learning and Convergence. International Conference of Knowledge
Management System, 7-9th July, Kuala Lumpur, Malaysia, 2005. This volume.
26. S.M.F.D Syed Mustapha. An Algorithm for Avoiding Paradoxical Argument among
the Multi-Agent in the Discourse Communicator. 8th International Conference on
Knowledge-based Intelligent Information & Engineering System, KES 04, New
Zealand, Vol I, pp350 – 356, 2004.