IKAT (Intelligent Knowledge Acquisition Tool) was created by Epistemics as a new knowledge toolkit similar to PCPACK but which is fully web-enabled. This work was undertaken as part of a DTI-funded Eureka project.
Need for IKAT - Knowledge Asset Management
Knowledge asset management concerns the need to assess the value of an organization's knowledge assets, against a backdrop that includes the organization's value chain, business strategy and market objectives. The aim of knowledge asset management is to characterize knowledge assets in terms of the agents (people and computer systems) who possess them, the tasks in which they are used, and their availability in terms of time, location, form and quality. Thus knowledge asset management provides a structured audit of an organization's knowledge assets that leads to many benefits associated with strategic decision-making, knowledge retention, corporate mergers, re-organizations, IT, training, etc.
Although the potential benefits are considerable, knowledge asset management is not easy - it can be very costly due to the complexities involved in acquiring, communicating and quantifying knowledge. Knowledge is notoriously difficult to acquire irrespective of whether it already exists or needs to be created from bodies of data. In addition, the difficulty of measuring and quantifying the value of knowledge makes it the ultimate intangible asset. Part of the solution is to undertake a knowledge audit in which the contribution of each knowledge asset is evaluated in terms of the organizational context - the goals, mission, value chain and strategy of the organization.
To do this requires a structured process of knowledge acquisition which minimises the impact on the business whilst maximising the quantity, quality, clarity and relevance of the acquired knowledge. The key to this process is the ability to efficiently capture knowledge from people. However, the intellectual assets possessed by people often comprises a large proportion of tacit knowledge - knowledge that people use in a 'sub-conscious' way to perform specialised tasks. As such, tacit knowledge is very difficult to capture and requires special methods developed by cognitive psychologists.
The current state-of-the-art in knowledge acquisition is the use of software tools whose functionality emulates that of special acquisition techniques. In many cases, the use of such software can increase the productivity and efficiency of the knowledge acquisition process. An important advantage of software-assisted knowledge acquisition is the ability to provide multiple visualizations of a single knowledge repository. Such visualizations allow different 'views' of knowledge which improves the efficiency of (1) further knowledge capture, (2) validation of knowledge, (3) publication and use of knowledge, and (4) future maintenance of knowledge. Other advantages of a software-assisted approach include automatic checking and validation of knowledge, automatic document and code generation facilities and improved transfer of knowledge structures to other IT systems.
For over ten years, Epistemics have been creating and using state-of-the-art software for knowledge acquisition. The new web-enabled toolkit, named IKAT, builds on the success of PCPACK, Epistemics' leading product for knowledge acquisition and knowledge-based system design which is used by world-leading organizations such as Airbus, Rolls-Royce, QinetiQ, Unilever and GKN. The new tool provides a set of web-enabled tools that are particularly focused on supporting the requirements of knowledge asset management. For example, they include tools to support special techniques such as laddering, concept sorting and repertory grid that allow tacit knowledge to be captured and explicitly represented.
IKAT is an important step forward in providing a cost-effective means for knowledge asset management. The IKAT toolkit provides access to a range of web-enabled knowledge acquisition tools that assist the knowledge engineer in characterizing the knowledge infrastructure of a particular problem domain. Each of the tools co-opts distinctive representational and visualization features with unique editing capabilities in order to provide an effective platform for the strategic management and valorization of corporate knowledge assets. The tools are as follows:
Laddered Grid Tool: The IKAT Laddered Grid Tool allows the knowledge engineer to construct laddered grids corresponding to decomposition hierarchies for a particular domain-specific relationship. The two most common types of hierarchies are taxonomic (is-a) and compositional (part-of); however, ladders can be constructed for any relationship (e.g. a functional decomposition), and mixed ladders (which feature multiple relation and element types) can also be used. The Laddered Grid Tool presents this hierarchical knowledge as a two-dimensional directional graph. The unique representational and visualization capabilities of the Laddered Grid Tool make it ideal for editing the hierarchical relationships between knowledge elements.
Repertory Grid Tool: The IKAT Repertory Grid Tool is an implementation of a psychological technique used for revealing hidden or implicit conceptual structures. In its basic form, the expert is asked to discriminate between triads of entities in a domain, by suggesting a construct that applies to two of the entities but not the third. The resulting construct is then applied to other entities in the domain. Random triads can be presented, and new constructs can be created to discriminate between them. The result is a series of constructs that can be used to describe the domain. The contrived format of the repertory-grid technique means that some of the constructs might not even be apparent to the expert. A variety of cluster analytic techniques can be applied to the completed repertory grid as the basis for revealing the relative similarity of elements and constructs. The hierarchical clustering of elements can reveal complex taxonomies that may not have been consciously accessible to an expert. The repertory grid technique has been used with considerable success in a variety of knowledge-rich domains. For example, in the field of medicine, clustering diseases, cures for diseases, or symptoms of diseases can lead to very useful taxonomies. In the field of psychiatry, the correct diagnosis of clusters of symptoms such as paranoia, schizophrenia, etc. is essential for the selection of successful therapeutic strategies. In archaeology, researchers have attempted to establish taxonomies of stone tools, funeral objects, etc. by applying repertory grid techniques. In general, whenever one needs to classify a "mountain" of information into manageable, meaningful taxonomies, the repertory grid technique of is of potential value. The technique can also reveal statistically significant dependencies between the characteristics of objects. In some cases, this is useful for formulating rules about the relationship between the properties of concepts, which itself is useful for determining the kind of inferences that can be derived from a body of data.
Card Sort Tool: The IKAT Card Sort Tool is a tool for sorting and categorizing groups of knowledge objects. The basic aim is to support the expert's grouping of concepts into significant clusters. The visual metaphor of the Card Sort Tool is that of taking a series of cards, each of which has a visual representation of a knowledge object on it, and then sorting them into piles to make distinctions between them. By doing a card sort, the various ways in which experts 'see' a set of concepts, and the various types of distinction that they make between them, can be made clear.
Protocol Editor Tool: The IKAT Protocol Editor is a tool to annotate knowledge-rich textual resources, otherwise known as protocols. Protocols are important sources for knowledge acquisition; they can include books or documents, manuals, or transcriptions of experts' commentaries or interviews. Phrases in the protocol can be marked up in the Protocol Editor using special marker 'pens'. These allow the user to note occurrences of, among other things, concepts, attributes, values, relations, processes or rules, or to make notes connected with particular phrases. The various phrases isolated during the marking up period can be used to develop a lexicon, or glossary, of important terms in the domain. The concepts, and so on, that are isolated by the editing process can also be used at an early stage in a project as raw material for other IKAT tool components.
Annotation Template Tool: The IKAT Annotation Template Tool allows a knowledge engineer to specify a template for a category of knowledge objects or a specific type of knowledge object, e.g. a specialization of a higher order concept. The template is designed using standard HTML technology and thus can support a potentially limitless variety of page layouts and information content. In addition, a number of application-specific formulas can be inserted into the template as the basis for automatically populating annotation pages with information about the inter-relationships between the annotated knowledge construct and other constructs in the knowledge project.
Annotation Tool: The IKAT Annotation Tool enables users to add annotations to their projects. An annotation page is essentially a standard HTML page that can be edited in a manner similar to the well-known graphical web authoring tools. Annotations can store explanatory text and multimedia information for any knowledge construct which features in the knowledge base. Since the annotation page is an HTML page it is easy to navigate to related knowledge objects using a native hyperlinking capability. Each annotation page is derived from an annotation template (see above), from which it inherits stylistic and layout information. The formulas defined in the underlying template are executed in the context of the annotation page in order to reflect the information infrastructure of a knowledge project. Such formulas are updated with every modification to the knowledge project, so the content of the annotation page always reflects the current project structure.
Diagram Template Tool: The IKAT Diagram Template Tool can be used to specify a template for a particular type of diagram that is subsequently created and used in the context of the Diagramming Tool (see below). The Diagram Template Tool allows a user to place restrictions on the types of knowledge objects that can feature in a diagram as well as specifying the type of diagrammatic notation to use in representing a knowledge object.
Diagramming Tool: The IKAT Diagram Tool allows the user to characterise and edit knowledge objects in a visually oriented manner using a predefined set of graphical modelling notations. Many modelling frameworks avail themselves of standard sets of graphical formalisms, which can be used to communicate modelling decisions in a concise and effective manner, e.g. the UML. The diagramming tool allows a user to specify any set of graphical notations, which characterise their preferred modelling methodology using the Diagram Template Editor (see above). Users can subsequently use these graphical modelling conventions to depict the inter-relationships between knowledge objects in a particular knowledge project. The Diagram Tool therefore enables end-users to exploit the full functionality of the IKAT Toolkit without renouncing their preferred modelling methodology. It also provides an easy to use graphical interface for describing the knowledge infrastructure of a target application domain.
Benefits of Using IKAT
The range of tools provided by the IKAT framework facilitates the process of eliciting, capturing and documenting knowledge. Each of the tools is specifically geared to representing knowledge in a particular way and is therefore differentially suited to the elicitation of different types of knowledge. Moreover, the graphical capabilities of each tool enables the knowledge engineer to visualize the knowledge infrastructure of a domain from a number of different perspectives and in a number of different ways. Collectively, the IKAT tools provide a rich set of resources designed to facilitate knowledge engineering and design activities. We regard knowledge engineering as essentially a knowledge-intensive design activity in which the competent selection and utilization of a variety of different software resources is central to problem solving success.
Other Products: PCPACK, MetaPACK