This paper proposes an initial design study to examine and test some of the key concepts and issues within a large-scale European research project that is exploring and aiming to realise learning as a process of knowledge maturing in the workplace. It will outline some of these concepts, based on a contemporary (or Web 2.0 driven) articulation of how ontologies can be acquired, externalised and exploited by a user-community and introduce a new role for learning dialogue - through developing work into „dialogue games‟. An initial scenario, or „thought experiment‟, is proposed that is grounded on currently available ontology development (SOBOLEO) and learning dialogue (InterLoc) web-technologies and how these could be integrated, or „mashed up‟, to improve the management, understanding and application of labour market information in the context of careers advice. Finally, we also consider the potential role of m-learning techniques and the implications about context that these give rise to.
the application of a DBR approach to develop digital dialogue games for learning
and intelligent dialogue systems [
The remainder of this paper will synthesise work in ontology maturing and learning dialogue to propose, or scope out, an initial design study for a system integration, or mashup, that exemplifies important aspects of learning as knowledge maturing and also considers m-learning and the implications this has for notions of context. This will be grounded on existing web-technologies (SOBOLEO and InterLoc) and a particular hypothetical user-scenario based on how practitioners giving careers advice to young people in a particular region of England can more effectively research and use Labour Market Information (hereafter LMI). We will introduce each of these technologies and their rationale before articulating their potential combination to realise knowledge maturing in careers advice settings. Note that this study is part of a federated set of studies that are examining different aspects of the (large-scale) knowledge maturing enterprise, where our particular emphasis is on: exploring and developing the relationship between notions of social learning and knowledge maturing; testing the technical integration issues related to the creation of suitable mashups; and, exploring the role of collaborative dialogue in the continuous development of knowledge networks. So this complements other ongoing design studies, such as those giving greater emphasis to the role of direct user and community engagement in the specification of system requirements and the design process in general.
Some important work that reconciles more (traditional) formal approaches to
ontology development with more participative (Web 2.0 driven) approaches has
been proposed by [
The starting point of their ontology maturing process model were the shortcomings
of the usual separation of creation and usage processes, performed by different sets
of people [
This led [
An ontology needs to be an appropriate representation of the domain with respect to the purpose of the ontologies required for a semantic application so that it is actually useful. That means, we need a quick, simple and work-integrated way to adapt and modify the ontologies. The aspect of social agreement requires that an ontology is a shared understanding of a given domain among all stakeholders. Therefore, the involved individuals deepen by and by their understanding of the real world and of an (appropriate) vocabulary to describe it. The development of an ontology underlies a process of continuous evolution where different levels of formality might co-exist within one ontology. The outcome is an adequate level of formality in the ontology, avoiding both overformalisation and the inability to apply semantic algorithms. This balance comes up in a continuous social learning process tightly coupled with the usage processes of the ontology.
The ontology maturing process model operationalizes this view and structures the ontology development process into four phases. Starting with simple tags, each user shall contribute to the collaborative development of ontologies. Thus, each community member can contribute new ideas (tags) emerging from the usage to the development of ontologies (phase I “Emergence of Ideas”). The community picks them up, consolidates, and refines them (phase II “Consolidation in Communities”) and formalizes with semantic relations towards lightweight ontologies (phase III “Formalization”) or even adds axioms (maybe with support of knowledge engineers) for improving inferencing processes (phase IV “Axiomatization”). In this way the users themselves can directly execute changes if needed. The time and costconsuming separation of ontology creation and usage is overcome.
The SOBOLEO tool (below) realizes this ontology maturing process model by
offering an easy-to-use interface (to allow the usage of semantic technologies also
for “ordinary” people) and ontology development integrated into the actual usage
processes; i.e. the semantic annotation and retrieval of web resources.
3.1. SOBOLEO
SOBOLEO [
In this way, new concept ideas are seamlessly gathered when occurring and existing ones are refined or corrected. The users can structure the concepts with hierarchical relations (broader and narrower) or indicate that they are “related”. These relations are also considered by the semantic search engine. That means, the users can improve the retrieval of their annotated web resources by adding and refining ontology structures.
So, whilst the role for and value of SOBOLEO is clearly argued above, it seems
important and enticing to ask whether the introduction of a specially designed
dialogue will improve the way in which its ontologies are developed, refined,
shared, used and generally understood. Or similarly, can we support specialised
dialogues that promote the phases of knowledge maturing described above? And
along similar lines, can we introduce learning dialogues to support collaborative and
social learning around ontology development and use? With this in mind we
consider work into digital dialogue games that is described below.
4. Learning Dialogues, Dialogue Games and Interloc
Our current dialogue game technology - InterLoc3 [
The interface in Figure 2 shows how each player participates in the dialogue game to produce a balanced, thoughtful, coherent and yet critical dialogue (about DNA testing in this case). It models natural dialogue through allowing players to either Contribute to the current state of the developing dialogue through selecting “Make Contribution” or Replying to a specific previous contribution by selecting “Reply”. Contributing to the dialogue places a response at the bottom of the display and Replying indents the responses below the specific contribution that is replied to - in a threaded way. This visual idiom contains affordances that achieve a balance of „keeping the dialogue moving forward‟ whilst allowing reflective asides and specific responses to previous contributions. All contributions or replies are made using these Move categories (Inform, Question, Challenge, etc.) and scaffolded through using specific Locution Openers (“I think…”, “I disagree because…”, “Let me elaborate…” etc.) that have to be used to perform the dialogue. Similarly, rules about the legitimate and logical responding openers, based on the specific openers that are replied to, are offered selectively - but these can be overridden to select the full range of options through selecting “More”. For example Figure 2 shows a player called Seb1 deciding to access the full range of moves and openers through selecting “More” instead of using “Because…” - which is the prompted response to the “Please give a reason…” opener. A model of turn-taking is also incorporated to ensure that the dialogues support: „listening‟ to others contributions; fairly balanced patterns of contribution; and, generally, the sort of coherent sequencing that results in reasoned discourses. An important point about the dialogue game approach is that all the textual contents contained in the Menus (i.e Move categories and specific locution openers) that realise the games are read in from xml files and so can be easily edited and 1 Note that the actual names have been anonymised but the gender retained. amended, to provide refined or completely new games. Ongoing work with early adopters is making this process even easier, through the development of a dialogue game-editing tool, which means that it would be possible to develop a knowledge maturing dialogue game (or KM-DG), or number of games, linked to ontology development and use through linking with, or „mashing up‟ with a technology such as SOBOLEO.
We now consider the benefits of „mashing up‟ the ontology development and
learning dialogue tools to investigate learning as knowledge maturing. Also as
evaluation results have shown [
In this section we consider how such a mashup, of ontology development and dialogue technologies, could be applied within a concrete knowledge maturing context. This is focussed on how Connexions Personal Advisers (P.A.s) located in one region of England use Labour Market Information (LMI) to advise young people. Connexions companies in England provide a range of services to young people aged 13 to 19, including careers guidance. The particular example used here is based on a fictional scenario that was developed by the Institute for Employment Research at the University of Warwick in conjunction with Connexions Kent. Specifically, it is based on how a P.A. might work with a young female interested in becoming a plumber. In this situation the P.A. would need to perform a number of knowledge maturing processes to research and mediate the LMI in a meaningful way for the young female. Namely, this would involve: 1. Aggregating and scaffolding; 2. Manipulating; 3. Analysing; 4. Storing; 5. Reflecting; 6. Presenting; 7 Representing; 8. Sharing; and, 9. Networking with other people. So below we make an initial attempt to summarise how the proposed mashup might assist with these processes. But to begin with it might be useful to sketch a conceptual overview of how this might work. Essentially, the P.A. is constructing a knowledge representation, or domain model, through continuous on-task activities (such as book marking through SOBOLEO) and through performing a dialogue with and about the developing ontology and its instantiation with LMI. This „engine‟ of a (continuously developing) knowledge representation plus various dialogue facilities can then support a range of features associated with the development and application of the knowledge for advising and problem solving. These features could be productively operationalised along the lines below.
Firstly, whilst they are researching suitable LMI, the P.A. could use a SOBOLEO type application to construct an ontology relevant to their situation and context in a continuous and embedded way, that is relevant to their particular domain (i.e. advising a female about becoming a plumber). This could be assisted through using location-based information to automatically direct the P.A.s search for LMI that is relevant to their locality. So in the first instance the P.A. could simply bookmark the resources without saying anything more specific about their nature or relationship to one another (i.e. simply „collect‟ LMI resources). Secondly, the P.A. will need to organise their information in a way that is meaningful to them, so they could perform a knowledge maturing dialogue game to create a more semantically rich and organised ontology through introducing categories and relations, such as “is similar to”, “is an example of”, “an exception is”, etc., to produce a personalised and semantically enhanced organisation of the LMI they have collected. They may even want to model „dialectical‟ relations, such as contradictions, inconsistencies and uncertainties. And through performing a knowledge maturing dialogue game to externalise these features and aspects they should be able to develop a better understanding, or informally learn, about the domain they are constructing (so these two processes are similar to maturing phase I). Thirdly, they may then want to collaboratively refine the ontology they have acquired and refined through dialogue games with colleagues, to produce a further refined and negotiated ontology, and in the process, again, they are likely to informally learn through these collaborative dialogues. So fourthly, they could then move to knowledge maturing processes that build on the existence of what is, now, more of an organisational or community ontology. This might involve other members of the organisation using the same technology to further develop, inspect and refine the ontology (so these third and fourth processes are similar to maturing phase II). Fifthly, once a relatively mature, negotiated and formal ontology has been produced (which is similar maturing phase III), it could then be compared with those produced for related knowledge domains, such as related parts of the Construction Industry in this case. These could then be shared with other P.A.s in other areas of the country who are also working with individuals expressing similar interests, or may want to learn about successful outcomes from advisory sessions elsewhere. Indeed, locally produced ontologies could serve as case studies that could enable the proliferation and re-use of similar knowledge maturing processes across the country. Sixthly, matured ontologies could provide visualisations, descriptions or dialogues that are tailored to different audiences, to accommodate different views on the same knowledge maturing processes and structures, such as the viewpoints of the clients of the P.A.s. So, to summarise the above, through combining a flexible ontology development and learning dialogue technology, a range of knowledge maturing services could be provided that include: 1. knowledge acquisition; 2. personal knowledge refinement; 3. collaborative knowledge refinement and negotiation of meaning; 4. informal learning; 5. collaborative learning; 6. support for advising and problem solving; 7. reflection and meta-cognition about the domain and its application; 8. rerepresenting domain knowledge for different audiences and purposes. Or summarising all this, arguably, having the potential continuously to develop a personal and community Ontology combined with the means to have a specialised and scaffolded dialogue about it, will potentially make the domain more understandable and the application of the ontology more powerful.
The work proposed in this paper is in progress and necessitating the next stage of
actually developing the mashup and testing it in a suitable application area, and
would clearly benefit from further investigation and mapping of the knowledge
maturing phases (e.g. exemplified by SOBOLEO) against practical examples of
„knowledge practices‟ and how the inclusion of dialogue refines our understanding
of these phases. However another key dimension that will potentially be important
to such a study is the need to „go mobile‟ and the implications this has for context.
In the scenario referred to above, this was represented as „sharing‟, depending on the
precise circumstances of the context in which the P.A. was working. It would
typically involve sharing new knowledge and understanding with other practitioners
operating in different contexts or possibly „mobile learner-generated contexts‟ [
Acknowledgements The authors are grateful to all members of the FP7 MATURE Project (http://matureip.eu/en/start) who have contributed to this work and V. Zacharias and M. Sagar who developed SOBOLEO and InterLoc 3, respectively. 15. Cook, J., Bradley, C., Lance, J., Smith, C. and Haynes, R. (2007). Generating Learning Contexts with Mobile Devices. In Norbert Pachler (Ed.), Mobile Learning: Towards a Research Agenda, WLE Occasional Papers in Work-Based Learning 1, London. Download from: www.wlecentre.ac.uk/cms/files/occasionalpapers/mobilelearning_pachler2007.pdf 16. Dourish, P. (2004) „What we talk about when we talk about context.‟ In Personal and Ubiquitous Computing 8(1), pp. 19-30