=Paper=
{{Paper
|id=Vol-34/paper-6
|storemode=property
|title=A Scenarios Mediated Approach for Tacit Knowledge Acquisition and Crystallisation: Towards Higher Return-On-Knowledge and Experience
|pdfUrl=https://ceur-ws.org/Vol-34/cheah_abidi.pdf
|volume=Vol-34
|dblpUrl=https://dblp.org/rec/conf/pakm/Yu-NA00
}}
==A Scenarios Mediated Approach for Tacit Knowledge Acquisition and Crystallisation: Towards Higher Return-On-Knowledge and Experience==
https://ceur-ws.org/Vol-34/cheah_abidi.pdf
A Scenarios Mediated Approach for Tacit Knowledge Acquisition and
Crystallisation: Towards Higher Return-On-Knowledge and
Experience
Cheah Yu-N Syed Sibte Raza Abidi
School of Computer Sciences School of Computer Sciences
Universiti Sains Malaysia Universiti Sains Malaysia
11800 Penang, Malaysia 11800 Penang, Malaysia
yncheah@cs.usm.my sraza@cs.usm.my
Capital of an organisation. Put simply, Human Capital is
the knowledge, skills, experiences and intuitions possessed
by individuals in an organisation; Structural Capital refers
Abstract to knowledge that is embedded within an organisation
The ‘Knowledge Age’ has fuelled the need to operational system, i.e. an organisation’s processes, work-
capitalise on organisation-wide Intellectual flows, systems, policies and procedures; and finally Social
Capital with the aim of gaining competitive Capital refers to the organisation’s relationships with its
advantage vis-à-vis a higher return-on-knowledge network of customers as well as its network of strategic
and experience. In this paper, we propose a novel partners and stakeholders. Of the three facets of
tacit knowledge acquisition and representation Intellectual Capital, in this paper we will focus on Human
strategy using Scenarios based on the assumption Capital.
that tacit knowledge can best be explicated Vis-à-vis Human Capital, the knowledge possessed by
through controlled challenge situations. We also an individual can be broadly differentiated between
describe in detail, a knowledge crystallisation Explicit Knowledge and Tacit Knowledge. Explicit
algorithm that is used to refine the acquired tacit Knowledge can best be described as canonical knowledge,
knowledge by modelling the natural mechanics of i.e. knowledge formalised within databases, business rules,
crystallisation and annealing. We conclude by manuals, protocols and procedures and so on. Tacit
asserting that our approach would further Knowledge is non-articulated or non-canonical
enhance organisation-wide performance through knowledge, i.e. knowledge that does not manifest as rules.
its superior quality and value-added delivery of Rather it exists as the domain experts’ skills, common
organisational services. sense and intuitive judgement whilst solving problems
[Che00]. The problem in many organisations today, we
believe, is that explicit knowledge has been given more
1 Introduction prominence and that experience- and skill-rich tacit
knowledge is ineffectively or even hardly captured and
In today’s ‘Knowledge Age’, there is an ever increasing
utilised. This reliance on explicit knowledge may also
demand for more sophisticated Knowledge Management result in rigid policies and thinking patterns that would
(KM) methodologies and techniques to provide the so-
hinder an organisation’s ability to gain competitive
called ‘next-generation business solutions’—solutions that
advantage and, thus, resulting in low return-on-knowledge
endeavour to provide a higher return-on-knowledge to the
and experience.
parent enterprise [Lan99]. It is widely acknowledged that
In this paper, we will focus on knowledge creation
an organisation’s competitive advantage and its capacity to
strategies [Non94], in particular the acquisition,
achieve higher return-on-knowledge is derived from the representation and crystallisation of tacit knowledge with
effective operationalisation and management of its
the aim of allowing tacit knowledge to be utilised
Intellectual Capital. With definitions abound, Intellectual
effectively to gain competitive advantage with higher
Capital is generally described as comprising the Human
return-on-knowledge and experience. By using healthcare
Capital, Structural Capital, and Relational (or Social)
as an exemplar domain, we present a novel tacit
knowledge explication approach that purports the
The copyright of this paper belongs to the paper’s authors. Permission to copy
presentation of Scenarios pertaining to hypothetical
without fee all or part of this material is granted provided that the copies are not
made or distributed for direct commercial advantage.
problem situations to healthcare experts and in turn to
Proc. of the Third Int. Conf. on Practical Aspects of record their ‘tacit’ problem-solving methodology and
Knowledge Management (PAKM2000) knowledge in solving the given problems. The specialised
knowledge extracting ‘scenarios’ are custom designed to
Basel, Switzerland, 30-31 Oct. 2000, (U. Reimer, ed.)
reflect atypical problems – i.e. not the kind of problems
http://sunsite.informatik.rwth-aachen.de/Publications/CEUR-WS/Vol-34/ that can be solved by routine procedures. Rather, these are
Y.-N. Cheah, S.S.R. Abidi 5-1
problems whose solutions may demand an interplay of provoking domain experts to act and apply their
informal and ad hoc intuitive (or based on experience) knowledge and skills to solve novel or atypical problems.
judgements with formal problem-solving strategy. Such a provocation is to be achieved by repetitively
The forthcoming discussion spans across the lifecycle presenting domain experts ‘hypothetical’ Scenarios
of knowledge creation; initiating with a discussion on the [Che00] pertaining to novel or atypical problems and then
proposed specialised knowledge extracting Scenarios vis- observe and analyse the domain expert’s tacit knowledge-
à-vis mental models (as suggested in cognitive science based problem-solving methodology and procedures. In
literature). We will discuss in detail, a scenario structure this context, the proposed problem-specific scenario
and representation scheme that makes reference to the presents domain experts the implicit opportunity to
notion of Meta-Scenario and its components. Next, we introspect their expertise and knowledge in order to
will proceed to discuss scenario acquisition mechanisms. address the given problem, to explore their ‘mental
Finally, we will present the latest extension of our work, models’ pertaining to the problem situation and solution,
i.e. a technique for incremental knowledge refinement and and finally to apply their skills and intuitive decision
categorisation through knowledge crystallisation which is making capabilities. This sequence, allows tacit
based on the novel notion of Knowledge Nucleation and knowledge to be ‘challenged’, explicated, captured and
Growth [Che00] – the formation of knowledge crystals by finally to be stored. This strategy is in line with the notion
the amalgamation of multiple contextually/structurally of ‘contrived’ knowledge acquisition techniques [Spe99,
similar scenario items. The work reported here purports a Sha90].
synergy between artificial intelligence techniques (for The novelty of our approach draws from the way
representation, reasoning and learning purposes) with ‘hypothetical’ atypical problem situation (a problem-rich
existing concepts and practices in knowledge scenario that characterises the various elements and
management. In conclusion, we assert that our approach peculiarities of a possible real-life situation) is derived.
would further enhance organisation-wide performance Another novelty stems from the goal of the our strategy –
through its superior quality and value-added delivery of i.e. to explicate tacit knowledge vis-à-vis charting out the
organisational services. domain expert’s thought processes, or more appropriately,
mental models (as in cognitive science) for a set of well-
defined problems. Having introduced and discussed the
2 Tacit Knowledge Acquisition and rational behind the use of scenarios, we will now provide a
Representation Using Scenarios description of scenarios.
Generally, traditional approaches, such as the interviewing
of domain experts, and subsequently observing and 2.1 Description of Scenarios
analysing their problem solving methodologies [Mor91], A scenario, by itself, is a customised, goal-oriented
obtaining knowledge from reference materials and narration or description of a situation, with a mention of
databases, and other techniques such as role playing, actors, events, outputs and environmental parameters. Put
talkback, 20 questions, repertory grid, etc. only work well simply, a scenario (a) depicts a sequence of distinct
to procure explicit knowledge. With this in mind, it is actions that might be taken to accomplish a particular task;
widely contended that traditional strategies are ineffective and (b) details the sequence of interactions – comprising
when it comes to acquiring an expert’s tacit knowledge. exchange of messages and responses to intermediate
This is because traditional knowledge acquisition outcomes – performed or experienced by entities to fulfil
techniques do not take into account the intrinsic origin and the goal [Bee98]. In essence, a scenario is a collection or
composition of tacit knowledge during its acquisition. sequence of hypothetical (but mimicking real) situations
In view that tacit knowledge is highly intangible, encountered by a domain expert, together with the
abstract and hidden, one can attribute its origin to be intermediate responses/actions by the domain expert and
seemingly incorporated, embedded or interleaved with henceforth describes one or more episodes or events of the
certain innate and essential skills – problem-solving skills, scenario.
analytical skills and generalisation or abstraction skills. From a cognitive science perspective, a scenario can be
We argue that it is the selective and systematic deemed as a means to explicate the domain expert’s
manipulation of these innate skills, subject to the nature mental model of the problem and its solution. Mental
and specification of the problem to be solved, that brings models are not explicit and need to be inferred from by
into relief the so-called tacit knowledge that we seek to putting them into action. This description of mental
capture. Furthermore, with regards to representation models therefore relates closely to the qualities of tacit
schemes, the mention of schemata and mental models vis- knowledge and its acquisition through the use of scenarios
à-vis tacit knowledge representation is of relevance here as [Far88].
it posits that mental models are perceptions of the world. From an artificial intelligence perspective, a scenario is
Thereby, an insight into tacit knowledge representation very similar to a Case. However, the major distinction
(within our minds) can be achieved by understanding the between the two is that a case is a real-life situation-action
cognitive make-up of such mental models. structure, whereas a scenario represents a sequence of
In order to acquire tacit knowledge from domain hypothetical situations carefully designed to draw out tacit
experts, we propose a strategy that is grounded in the knowledge. Furthermore, cases are merely ‘frozen’
assumption that tacit knowledge can best be explicated by snapshots of an episode with no apparent attempt to
Y.-N. Cheah, S.S.R. Abidi 5-2
capture significant temporal or sequential elements. On the situational descriptions. Since tacit knowledge is both rich
contrary, as per our specifications, scenarios can manifest in content and is also ‘deep rooted’, such contextual
a temporal nature whereby they can capture the sequence information is imperative for the successful illustration of
of events as they may have occurred during a particular the bigger picture, i.e. the sequence of episodes and
episode. In summary, we posit that a scenario is seemingly events. To support our claim pertaining to the efficacy of a
a more apt representation of an episode or situation, as scenario-based representation, we will now discuss in
compared to the traditional case-based representation. detail the proposed generic structure of a scenario.
Functionally, the proposed scenarios is akin to the
Scripts knowledge representation construct [Sch77],
However, we would like to point out that scripts are 3 Generic Scenario Structure and
somewhat limited in representing the context of the Representation: An Overview
situation in its entirety. On the contrary, our proposed Scenarios may be composed of four main components
scenarios cater for the clear representation of context vis- [Sch97, Pot94]: Meta-Scenario, Scenario-Construct,
à-vis the provision for the linking of contextual documents Episode and Event. For pragmatic reasons, scenarios are
to the sequence of events (described within the scenario). represented by a four-tier scheme where Meta-Scenarios
We believe that the explicit characterisation of the full are placed at the top level followed by Scenario-
context – i.e. a description of the social settings, resources Constructs, Episodes and Events at the bottom level (see
and goals of the users [Nar92] – in describing a particular Figure 1).
situation is an important consideration with regards to
META-SCENARIO
Class ID Class Name Sub-Class List (1 to n)
SCENARIO-
CONSTRUCT
ID / Description / Trigger Event Episode List (1 to n) Concluding Event Crystallisation
Context / Timestamps Factor
EPISODE
Episode ID Episode Descr. Event List (1 to n)
EVENT
Event ID Event Type Actor Object Parameter-Value List (1 to n)
Figure 1: The Scenario Structure outline.
SCENARIO-CONSTRUCT
Scenario ID: 990713.1520
Scenario Description: First-aid CPR on adult male, 57 years of age. Bystander present.
Location: Roadside
Contextual Link: CPR, Elderly adult, First-aid
Start Timestamp: 1520
End Timestamp: 1538
Trigger Event: EV0001
Episode List Elements: EP0001, EP0002, EP0003, EP0004, EP0005
Concluding Event: EV0016
Crystallisation Factor: 24
PARAMETER-VALUE
EPISODE EVENT LIST ELEMENT
Episode ID: EP0001 Event ID: EV0002 Parameter-Value ID: PV0002
Episode Description: Assessment Event Type: Action Parameter: Shake
Event List: EV0002, EV0003, Actor: First-aider Value: Shoulder of Patient
EV0004, EV0005 Object: Patient
Parameter-Value List: PV0002
Figure 2: Sample Scenario-Construct, Episode, Event and Parameter-Value List Element.
Y.-N. Cheah, S.S.R. Abidi 5-3
The scenario storage medium – i.e. a Scenario Base – A unique feature of the Scenario-Construct is the
adheres to the same scheme such that the various Contextual Link field, which stores keywords to help
components of a scenario are stored in distinct locate (through a search on specific document bases)
repositories. An exemplar Scenario-Construct with a formal or informal documents containing contextual
sample Episode, Event and Parameter-Value List Element information pertaining to the episodes and events of a
from a cardiopulmonary resuscitation scenario base are particular scenario.
shown in Figure 2. We discuss, below, the four scenario The Scenario-Construct also has a Crystallisation
components. Factor field that indicates how often the scenario was
accessed and judge as useful. The role of this field will be
3.1 The Meta-Scenario Component further discussed in Section 5.2.
The Meta-Scenario component serves to implement a two- 3.3 The Episode Component
level (class and sub-class) categorisation of scenarios.
Each category is called a class of scenarios and would The Episode component stores details of individual
have a series of Sub-Class List Element (one for each sub- episodes of a scenario (see Figure 5). Each Episode
class). Each meta-scenario could have the representation comprises an Event List that stores the sequence of events
shown in Figure 3. that make up an episode in a scenario.
Example Example
Class ID CL0001 Episode ID EP0001
Class Name CPR Episode Description Assessment
Scenario Sub-Class CPR for Adult Event List EV0002,
Scenario List 990713.1520, EV0003,
990726.2053 EV0004,
EV0005
.
.
. Figure 5: Representation of an Episode.
Scenario Sub-Class CPR for Infant
Scenario List 990804.1037 3.4 The Event Component
The Event component stores details about individual
Figure 3: Representation of a Class in a Meta-Scenario. events. The representation for an Event is shown in Figure
Shaded rows indicate a Sub-Class List Element. 6. There are three Event Types: Normative – events that
are expected to occur on a normal basis, Obstacle – events
3.2 The Scenario-Construct Component that hinder the progress of the task, and Action – events
The Scenario-Construct – a constituent of a scenario – that define the course of action undertaken by an actor.
stores the description of individual scenarios. The
Example
representation for the Scenario-Construct is shown in Event ID EV0002
Figure 4. Scenario-Constructs comprise a sequence of Event Type Action
episodes that are arranged in chronological order to mimic (Actor) First-aider
the temporal characteristics of the scenario. Such a (Object) Patient
representation scheme ensures tractability in terms of the Parameter-Value List PV0002
sequencing (or chaining) of multiple episodes within a
scenario. Figure 6: Representation of an Event.
Example The IDs of parameters and values of an event (in the
Scenario ID 990713.1520 form of Parameter-Value List Elements) are stored in the
Scenario First-aid CPR on adult male, Parameter-Value List.
Description 57 years of age. Bystander
present. Location: Roadside.
In our scenario structure, episodes and events are
Contextual Link CPR, Elderly adult, First- generic in nature until they are bound at the Scenario-
aid. Construct (scenario base) level through the order in which
Start Timestamp 1520 they are arranged and also through the effect of the Start
End Timestamp 1538 and End Timestamps. Having formalised the scenario
Trigger Event EV0001 representation, we would now discuss the scenario-based
Episode List EP0001, EP0002, EP0003,
EP0004, EP0005
(tacit) knowledge acquisition methodology.
Concluding Event EV0016
Crystallisation 24
Factor 4 Tacit Knowledge Acquisition
Methodology: The Use of Scenarios
Figure 4: Representation of a Scenario-Construct. To facilitate organisational knowledge acquisition
activities, we have developed a tool called the Scenario
Composer [Che00] that facilitates domain experts to
Y.-N. Cheah, S.S.R. Abidi 5-4
respond to a given scenario through the use of a series of derived from existing solved scenarios by way of
electronic forms whose attributes correspond to a selecting a Point of Interrogation (POI) – a distinct
particular component of a scenario, i.e. Meta-Scenarios, point in the scenario between two events of type
Scenario-Constructs, Episodes and Events. They prompt Obstacle or Normative and an event of type Action.
domain experts to provide information or suggest values to The result is a challenge scenario that is then
the various scenario-defining attributes presented in the presented to the domain expert for the explication of
electronic form. Figures 7 and 8 are screenshots of the his/her tacit knowledge (see Figure 9). The construct
Scenario Composer following the Challenge and POI captures the domain
expert’s response, i.e. the explicated tacit knowledge.
For practical purposes, once a scenario base is
sufficiently populated with knowledge – derived from both
solved and challenge scenarios – it can then be used for an
assortment of knowledge-driven activities. In the
healthcare domain, for instance, it can be used for
Healthcare Enterprise Modelling [CA99, Che99b].
As it turns out, there is no restriction on the terms used
by the experts. Therefore, there is a need to integrate
domain ontologies with the Scenario Composer to enforce
a certain degree of standardisation. Without doubt, domain
ontologies hold an important place in our scenario-based
tacit knowledge acquisition mechanism. This is because
experts tend to use different, though similar, terms in
expressing themselves. The inconsistent use of terms
could cause serious problems especially when scenarios
are eventually used for inferencing purposes.
Figure 7: Scenario-Construct screenshot. Ontologies, in our scenario context, could be integrated
or implemented in a pre- or post-input fashion. We can
also view it as being proactive or reactive. This means that
we could perform standardisation either before or after a
scenario is accepted into the scenario base. Ideally, this
should be done before a scenario is accepted to avoid
ambiguity in the event the expert is no longer unavailable
to provide clarification. Upon detecting a potentially
ambiguous terminology (such as those that are applicable
in more than one context, or those that have more than one
meaning), this pre-input mechanism would suggest
standardised terms for the expert to choose. In the post-
input version, the expert has little or even no control of
how this mechanism would standardise the terms used.
The process of acquiring tacit knowledge and making
it explicit is only a small step in the ongoing efforts to
‘create’ knowledge [Non94]. Moreover, the scenario base
in its ‘natural’ state is deemed inefficient in view that the
Figure 8: Episode, Event and Parameter-Value List scenario items are not categorised in such a way that
Element screenshot. would guide potential inferences by limiting an inference
engine’s search scope. Therefore, in the following
Our novel scenario acquisition exercise distinguishes sections, we would explore further on the modes and
between two types of scenarios: phases of knowledge creation leading to the refinement
1. Solved Scenarios: Scenarios that define actual and categorisation of the scenario base through the
situations/problems that have already been crystallisation of the explicated tacit knowledge.
encountered and solved/addressed by domain experts.
They are akin to traditional form-based cases that are
acquired through traditional knowledge acquisition 5 Knowledge Crystallisation
techniques. Scenario bases start out as having only
solved scenarios. Nonaka proposes four processes or modes that effectuate
2. Challenge Scenarios: Scenarios that represent atypical the transformation between tacit knowledge and explicit
situations and are posed to domain experts as a knowledge [Non94]. These transformations span across
challenge to their expertise. Challenge scenarios are in five distinct phases, as shown in Figure 10. In a KM
line with our contention that tacit knowledge is best parlance, crystallisation is an integral process in the
explicated when experts are required to solve atypical creating concepts phase—it refers to the process where
problems. In most instances, challenge scenarios are
Y.-N. Cheah, S.S.R. Abidi 5-5
Event ID Event Type Event Description
Scenario Trigger EV0001 Obstacle Patient has pain at centre
Event of chest, lasting more than
First-aid a few minutes, radiating to
CPR on adult shoulders, neck and arms.
male, Episode EV0002 Action First-aider shakes shoulder
57 years of of patient gently and shout Challenge
age. to ask if patient is
Bystander alright.
Present EV0003 Obstacle Patient’s state of
consciousness is
unresponsive. POI
Action First-aider calls for help.
Action First-aider requests
bystander to telephone
Emergency Medical Services.
Action Place patient in a
comfortable position. Expert’s
. . . . Response
. . . .
. . . . +
Concluding Normative Patient’s pulse is 83 beats ‘Tacit
Event per minute and breathing at Knowledge’
15 breaths per minutes.
Emergency Medical Service
arrives 23 minutes after
call made by bystander.
Figure 9: A portion of a Challenge Scenario showing the Challenge, Point of Interrogation (POI) and the Domain
Expert’s Response.
Socialisation Externalisation Internalisation Combination
Sharing Cross-
Tacit Creating Justifying Building an levelling
Knowledge Concepts Concepts Archetype Knowledge
KNOWLEDGE BASE
Tacit Conceptualisation Explicit
Knowledge/ Knowledge/ Crystallisation
Perspectives & Externalisation Concepts
Figure 10: The Five Phase Model of the Organisational Knowledge Creation Process [Non95].
the various departments in an organisation, test the reality the satisfaction of mutual constraints and unification
and applicability of the created concepts. As a amongst an ensemble of knowledge items, akin to the
consequence, knowledge that is proven effective, useful process of arrangement of ions, as per chemical rules, to
and objective is maintained and perpetuated. Now, from a form a crystal. We call the knowledge unit created by our
chemical parlance, crystallisation is interpreted to mean to crystallisation process as a Knowledge Crystal.
‘solidify and internally arrange’. A knowledge crystal can be viewed as a structure with
In our work we intend to map the notion of chemical a repetitive arrangement of scenario items in various
crystallisation to a knowledge creation context, whereby: perspectives, with the objective refine the scenario base at
(a) knowledge items (scenario items in our case) mimic the Scenario-Construct level. Basically, there can be two
molecules, ions or atoms in a supersaturated chemical stages in knowledge crystal formation:
solution; and (b) the creation of concepts is achieved by
Y.-N. Cheah, S.S.R. Abidi 5-6
1. Nucleation: The formation of a new child knowledge Nucleation could be initiated when needed by the
base in a heterogeneous (complex) knowledge base. scenario base administrator or through an automated
2. Growth: The repetitive addition of free scenario items. seeding process. This automated process follows an
A prerequisite for both processes is supersaturation of analysis of the knowledge base’s content and nucleation
scenario items. This means that crystallisation could only occurs as often as needed depending on the requirements
occur when the knowledge density of a knowledge base in creating child knowledge bases.
exceeds a predefined level [Bun96].
5.2 Growth
5.1 Nucleation
Growth occurs after nucleation is established. It is a
Nucleation begins with the presence of Knowledge Seeds repetitive addition of scenario items to the nuclei
in the knowledge base and they are analogous to (knowledge seed) through the formation of links to the
impurities in a chemical solution. In the state of nuclei. Depending on the knowledge seed, these scenario
supersaturation, these knowledge seeds form the nuclei for items can be of the same structure and/or the same context.
the growth of knowledge crystals. Conceptually, Ideally, these links should be established between
knowledge seeds serve as a platform on which scenario scenario items of similar structure and context, i.e.
items grow on. originating from a combination knowledge seed. Not only
There are three types of knowledge seeds: would this fit closely to the original definition of chemical
1. Structural: The structural knowledge seed ensures that crystallisation but it would also facilitate the
only scenario items of the same structure may implementation and operationalisation of emergent child
interlink to it. The scenario base could potentially knowledge bases. Child knowledge bases, comprising
contain Scenarios, Episodes and Events that have scenario items of similar structure and context, can
different structural composition. therefore be envisaged as being more subject-focused in
2. Contextual: The contextual knowledge seed ensures terms of their application domain and are also easier to
that only scenario items of the same context may manipulate in terms of query construction.
interlink to it. It forms a knowledge crystal that is Besides the knowledge seeds’ structural and contextual
contextually uniform. properties, we posit that the scenario items’ ‘temperature’
3. Combination (Structural and Contextual): This or energy level (akin to that in simulated annealing
knowledge seed ensures that only scenario items of [Tay99, Kir83]) is an important factor in the growth stage.
the same structure and context may interlink to it. The notion of the energy of a scenario item could be
The structure of a knowledge seed may consist of the translated into the amount of activity surrounding a
following items (see Figure 11): particular scenario item, i.e. the number of times it was
1. Seed ID: States the identification number of the referred or used by the system. We propose that when a
knowledge seed. scenario item is referred to and is frequently judged as
2. Seed Type: Determines if the seed is a Contextual or useful, its accessibility is said to increase. Thus, its level of
Structural knowledge seed or a Combination. energy (or activity) decreases, i.e. for a scenario item to be
3. Context: States the context of the knowledge seed. easily accessed, it should not ‘move’ too much. When the
This will be used as the basis to attract scenario items energy level of the scenario item decreases, its ability to
and ensures that the knowledge crystal is contextually crystallise or link to other scenario item increases (in line
uniform. with thermodynamic principles). Conversely, when a
4. Scenario Structure: This field lists down the attributes scenario item is referred to and is frequently judge as not
(fields) of scenario items that can be bound to the useful, its accessibility decreases and energy increases.
Knowledge Seed. This ensures that only scenario When this occur, it is less likely to bind with other
items that are structurally the same can link to the scenario items. These relations are summarised in Table 1.
seed (as per Structural Knowledge Seeds). It should also be noted that a scenario item can only
5. Scenario Item List: This field lists down the Scenario crystallise when its energy level drops below a certain
IDs of all Scenario-Constructs that is linked to the level, i.e. it is accessed and judged as useful frequently
seed, i.e. crystallised. enough. Therefore, a scenario item’s frequency of ‘useful’
access would need to exceed a predefined threshold before
Example it can crystallise.
Seed ID KS0001 In order to facilitate the crystallisation process, a
Seed Type Contextual
Crystallisation Factor (CF) field is added to the Scenario-
Context First-aid
Scenario [“Scenario ID”, “Scenario Construct representation (as mentioned in Section 3.2) to
Structure Description”, “Contextual Link”, store the number of times the scenario item was accessed
“Start Timestamp”, “End Timestamp”, and judged as useful. The CF would be decremented when
“Trigger Event”, “Episode List”, it is not judged as useful after a specified period of time. It
“Concluding Event”] is possible for the CF to have a negative value in the event
Scenario [“s.19990713.1520”, ...]
Item List that it is not useful for a long time.
Upon completing the crystallisation process, the
Figure 11: Structure of a Knowledge Seed. resultant scenario base may contain (free) scenario items
as well as knowledge crystals. Free scenario items are the
Y.-N. Cheah, S.S.R. Abidi 5-7
Frequency of Access Accessibility Energy Level Crystallisation
↑ ↑ ↓ ↑
↓ ↓ ↑ ↓
Table 1: The Relation between Frequency of Access, Accessibility and Energy Level in the process of Crystallisation.
items of the original scenario base that did not bind to any
of the formed crystals. In fact, the resultant scenario base 5.3 The Crystallisation Algorithm
changes from time to time depending on whether new
scenario items are added to the scenario base. Figure 12 In an attempt to crystallise the scenario base, we have
illustrates the states before and after (or during) incorporated into the Scenario Composer, a knowledge
crystallisation. crystallisation component. The steps executed by this
In the Figure 12, two crystals are formed with two component in the crystallisation process are as follows (A
different types of knowledge seeds. Scenario Item ‘E’ is flowchart of the algorithm is shown in Figure 13):
shown linking with the crystal on the left as they are of Step 1: Create Knowledge Seed: A knowledge seed is
similar context and/or structure. Scenario Item X (a free created by the domain expert or scenario base
scenario item) is on its own, as there are no crystals with a administrator using the representation detailed in
structure or context similar to its own. Section 5.1. For a newly created knowledge seed,
With reference to Figure 12, crystallisation is shown as the Scenario Item List is an empty list.
an on-going repetitive process, executed in parallel even Step 2: Embed Knowledge Seed: The knowledge seed is
while more knowledge is accessed, shared and added into ‘placed’ into a scenario base to initiate the
the scenario base. Functionally, crystallisation is crystallisation process.
terminated momentarily when the density of scenario Step 3: Attract Similar Knowledge Units: Search the
items falls below the predetermined threshold. scenario base for scenario items that have a similar
context and/or structure as the seed (depending on
the Seed Type field).
BEFORE
SCENARIO BASE
Scenario Item 1
Scenario Item E
Scenario Item 3
Scenario Item X
Scenario Item A
Scenario Item 4 Scenario Item D
Scenario Item 2
Scenario Item C
Scenario Item B
CRYSTALLISATION
AFTER / DURING
SCENARIO BASE
Scenario Item E Scenario Item 3
Scenario Item 1
Scenario Item A Scenario Item B
Scenario Item 2
Scenario Item C Scenario Item D
Scenario Item 4
Scenario Item X
Knowledge Crystal/
= Sub- or Child Scenario Base = Knowledge Seed
Figure 12: The Scenario Base Before and After (or During) Crystallisation.
Y.-N. Cheah, S.S.R. Abidi 5-8
Step 4: Select Most Similar Knowledge Unit: Among the then added en bloc to the seed’s Scenario Item List.
matched items, select the scenario item that has However, we choose not to do this (at least for the time
been frequently accessed and judged as being the being) to mimic more closely the crystallisation process in
most useful, i.e. having the highest CF. chemistry (which is incremental in nature). The
Step 5: Associate Knowledge Unit to Knowledge Seed: incremental approach is also more time consuming.
Link the selected scenario item to the knowledge Nevertheless, from a chemistry point of view, the slower
seed by way of adding the scenario item’s the crystallisation process, the better the quality of the
Scenario ID to the seed’s Scenario Item List. crystals formed. However, at the moment we are not able
Step 6: Prioritise Selected Knowledge Units: Rearrange to demonstrate any clear advantage for our choice.
the Scenario Item List (of the seed) so that the CF Step 6 allows the Scenario-Construct with the highest
of the scenario items is in descending order. CF to be at the head of the Scenario Item List. This
Note that crystallisation begins after Step 2, provided increases the efficiency of potential inferencing strategies
the scenario base is sufficiently populated by free scenario as the most reliable or useful scenarios are considered
item. A user-defined threshold determines the number of first. In the previous section, we proposed that the CF of
free scenario items that are needed to initiate the the Scenario-Construct will be decremented, after a
crystallisation process. predetermined period of time, if it is deemed not useful.
In Step 4, as mentioned before, the selected scenario This strategy ensures that outdated or less useful scenario
item must have a CF that exceeds a predefined threshold items would eventually be among the last to be considered
before it can be linked to the knowledge seed in Step 5. during inferencing.
In Steps 4 and 5, the matched items could have simply
been sorted according to their CF in descending order and
Start
Step 1 Create Knowledge Seed
Step 2 Select Scenario Base
No Free Item
Wait count exceeds
threshold
Yes
Search Scenario Base for Scenario items
Step 3 with same context and/or same
structure as Knowledge Seed.
Among matched items, select Scenario
Step 4 item that has the highest CF.
No Crystallisation
Factor exceeds
threshold
Yes
Link Scenario item to Knowledge Seed
Step 5 by adding the Scenario item’s Scenario
ID to the seed’s Scenario Item List.
Rearrange Scenario Item List with the
Step 6 CF of Scenario items in descending
order.
Figure 13: The Knowledge Crystallisation Flowchart.
Y.-N. Cheah, S.S.R. Abidi 5-9
After Step 6, if the number of free scenario items still not useful. Garbage collecting can potentially be executed
exceeds the threshold, the process continues from Step 3. in parallel with the existing crystallisation algorithm after
The crystallisation process stops momentarily when the Step 3 where instead of selecting the most popular and
number of free scenario items drops below the threshold useful scenario item for crystallisation, we select the one
only to resume when the threshold is exceeded. that is least accessed and hence is removed from the
We argue that the manner in which crystallisation is scenario base. A detailed discussion of scenario-based
taking place by modelling chemical crystallisation does in garbage collecting is beyond the scope of this paper.
fact conform to Nonaka’s original view where the Nevertheless, we take this opportunity to highlight its
explicated tacit concepts are tested for reliability and practicality and possibility.
applicability [Non94]. When a user accesses and evaluates In general, we will like to point out that our approach
a particular scenario item, he/she is actually testing and for knowledge crystallisation renders a more physical and
affirming its applicability and usefulness. Therefore, we autonomous connotation such that the knowledge itself
argue that the more a scenario item is accessed and judged undergoes a proactive automatic categorisation as
as useful, the more applicable the scenario item is deemed compared to the original, more social and reactive
to be and more crystallised the concepts or scenarios. approach. Note that our approach to crystallisation aims to
The crystallisation algorithm can be adapted to link scenario items to a knowledge seed in a crystal-
perform garbage collecting functions, i.e. the removal of forming paradigm that ranks the scenario items in terms of
scenario items that prove to be very isolated cases and are their applicability and usefulness.
CONTEXTUAL_LINK+, START_TIMESTAMP,
END_TIMESTAMP, TRIGGER, EPISODES+,
CONCLUDING, CRYS_FAC)>
]>
OBJECT?, PARAMS_VALUES+)>
#REQUIRED>
(normative|obstacle|action) #REQUIRED>
IDREF #REQUIRED>
]> ]>
Figure 14: Exemplar Scenario-Construct DTD.
First-aid CPR on adult male, Assessment
57 years of age. Bystander present.
Location: Roadside.CPRElderly
adultFirst-aid15201538First-aider24
Figure 15: Exemplar Scenario-Construct instance in XML.
Y.-N. Cheah, S.S.R. Abidi 5-10
5.4 The Final Representation of Scenarios
Ideally, the explicated and crystallised tacit knowledge References
would be stored in a representation language that is
portable, flexible and expressive. Currently, a likely [Bee98] G.W. Beeler, W. Rishel, A.M. Shakir and M.
choice of representation language is the Extensible Mark- Walker. Message Development Framework
up Language (XML). The choice of XML is justified Version 3.1. Health Level Seven, Inc., 1998.
further by the fact that it is easily parsed and thus, [Bun96] H.R. Bungay III. Crystallization.
facilitates the translation of the XML document into other http://www.esb.ucp.pt/~bungah/coag/ crys.htm,
representational language. Exemplar Document Type 1996.
Definition (DTD) fragments for the Scenario-Construct, [Che00] Y.-N. Cheah and S.S.R. Abidi. A Strategy for
Episode and Event representations and their XML Knowledge Acquisition and Representation: A
instances are shown in Figures 14 and 15 respectively. Case for Scenarios. 3rd Int. Conf. on The
Practical Application of Knowledge
Management (PAKeM 2000), Manchester,
6 Concluding Remarks 2000.
[Che99a] Y.-N. Cheah and S.S.R. Abidi. Healthcare
We believe that we have initiated a novel scenario- Knowledge Management Through Building
mediated approach to facilitate the capture and utilisation and Operationalising Healthcare Enterprise
of the domain experts’ tacit knowledge, thereby leading to Memory. In P. Kokol, B. Zupan, J. Stare, M.
the crystallisation of the acquired knowledge. Here we will Premik and R. Engelbrecht (Eds.), Medical
like to point out that we do understand that the basis of our Informatics Europe (MIE ’99), Ljubljana,
tacit knowledge acquisition approach is to some extent Slovenia, Amsterdam: IOS Press, 1999.
akin to traditional knowledge acquisition strategies, such [Che99b] Y.-N. Cheah and S.S.R. Abidi. Evaluating the
as critiquing, role playing and simulation. However, we Efficacy of Knowledge Management and
believe that the novelty of our approach derives from the Organisational Memory Toward Healthcare
following facts: (1) tacit knowledge is ‘invoked’ by not Enterprise Modelling. Workshop on Knowledge
mere interviewing the domain expert or document Management and Organizational Memories,
analysis, rather by subjecting the domain expert to solve 16th Int. Joint Conf. on Artificial Intelligence
‘controlled’ challenges (hypothetical’ atypical problem (IJCAI ’99), Stockholm, 1999.
situation) that itself are derived from existing solved real- [Far88] M.U. Farooq and W.D. Dominick. A Survey of
life scenarios; (2) the scenario knowledge construct Formal Tools and Models for Developing User
represents the hierarchical make-up of tacit knowledge in Interfaces. Int. Journal of Man-Machine
terms of an ensemble of distinct knowledge units; (3) the Studies, 29:479-496, 1988.
acquisition of tacit knowledge follows the formal [Kir83] S. Kirkpatrick, C.D. Gerlatt Jr. and M.P.
structural specification of the scenario, thereby ensuring a Vecchi. Optimization by Simulated Annealing.
mapping of user-mediated knowledge items (with varying Science, 220:671-680, 1983.
degrees of specificity and context) to a scenario structure; [Lan99] A. Lancini and E. Mercier-Laurent. Traffic
and (4) the crystallisation of acquired tacit knowledge in Road Accidents Return on Experience.
terms of the chemical crystallisation process. Discovering, Organizing and Sharing
Knowledge to Minimize Risks and Costs in a
In this paper, we have discussed the said tacit knowledge Mutual Insurance Company. 6th Int. Symp. on
acquisition and representation approach that attempts to the Management of Industrial and Corporate
capture the essence of expert-quality problem-solving Knowledge (ISMICK ’99), The Netherlands,
using scenarios. We have also seen how natural 1999.
phenomena, such as crystallisation and annealing, can be [Mor91] I. Morrison, B.A. Schaefer and B. Smith.
effectively adapted into Knowledge Management efforts Knowledge Acquisition: The Acquire®
to refine and categorise knowledge and to allow it to Approach. 1st Semi-Annual Conf. in Policy
dynamically evolve into scenario or knowledge bases that Making and Knowledge Systems, 1991.
are capable of providing relevant and up-to-date [Nar92] B.A. Nardi. The Use of Scenarios in Design.
knowledge on-demand. The definition and development of SIGCHI Bulletin, October, 1992.
scenarios, the crystallisation algorithm and the Scenario [Non94] I. Nonaka. A Dynamic Theory of
Composer are ongoing. Nevertheless, it is hoped that our Organizational Knowledge Creation.
methodology would lead to the improvement of Organization Science, 5(1):14-37, 1994.
organisation-wide return-on-knowledge and experience [Non95] I. Nonaka and H. Takeuchi. The Knowledge-
resulting in the superior quality and value-added delivery Creating Company. New York: Oxford
of organisational services. University Press, 1995.
[Pot94] C. Potts, K. Takahashi and A. Anton. Inquiry-
Based Scenario Analysis of System
Requirements. Int. Conf. on Requirements
Y.-N. Cheah, S.S.R. Abidi 5-11
Engineering (ICRE ’94), Colorado Springs,
1994.
[Sch77] R.C. Schank and R.P. Abelson. Scripts, Plans,
and Knowledge. In P.N. Johnson-Laird and
P.C. Wason (Eds.), Thinking: Readings in
Cognitive Science. Cambridge: Cambridge
University Press, 1977.
[Sch97] A.C. Schultz, J.J. Grefenstette and G.A. De
Jong. Learning to Break Things: Adaptive
Testing of Intelligent Controllers. In T. Baeck,
D.B. Fogel and Z. Michalewicz (Eds.),
Handbook of Evolutionary Computation,
Bristol: IOP Publishing Ltd., 1997.
[Sha90] N.R. Shadbolt and A.M. Burton. Knowledge
Elicitaion. In J.R. Wilson and E.N. Corlett
(Eds.), Evaluation of Human Work: A Practical
Ergonomics Methodology, 321-345, London:
Taylor and Francis, 1990.
[Spe99] P.-H. Speel, N. Shadbolt, W. de Vries, P. H.
van Dam and K. O’Hara. Knowledge Mapping
for Industrial Purposes. 12th Workshop on
Knowledge Acquisition, Modeling and
Management (KAW ’99), Banff, 1999.
[Tay99] T. Tay. Groping the Elephant.
http://www.comp.nus.edu.sg/~tayhanbo/
cs1305/index.html, 1999.
Y.-N. Cheah, S.S.R. Abidi 5-12