Knowledge Representation and Reasoning

Unit 5: Introduction to Modelling

Overview

As per the course website, "This week we will focus is on understanding the concept of modelling and knowledge engineering. The lecturecast will discuss the principle of knowledge engineering."

My Reflection

Overall Reflection

I found the topic of knowledge modelling interesting, just like the topics covered since the beginning of the module. In a way, knowledge modelling can be considered as a the group of techniques that allow knowledge representation. Therefore, I can see the topic as central to Knowledge representation, and I think it is the corner stone of the whole module.

The unit included a lecturecast, three reading materials and formative activities. The readings were quite intersting for me, especially Bimba et al. (2016), as shown below, as it gave an overview of the different knowledge models, in which knowledge can be represented to machines.

Reflection on Readings

The following is a summary of the three reading materials of the unit, as per my understanding:

Towards knowledge modeling and manipulation technologies: A survey: This paper by Bimna et al. (2016) categorises knowledge models into four main types: linguistic knowledge base, expert knowledge base, ontology, and cognitive knowledge base. Linuistic knowledge base is further categorised into FrameNet, WordNet, and ConceptNet. Expert knowledge base is sub-categorised into rule-based systems and fuzzy rule-based systems. Ontology is categorised into application, domain, generric and representation. Cognitive knowledge base is not further categorised. See the second artefact below for a mind map that I created to summarise these models. The paper also presents the limitations and used tools for each knowledge model.
On-To-Knowledge: Ontology-based Tools for Knowledge Management: This paper by Fensel et al. (2000) presents an EU-funded project that pioneered ontology-based tools for knowledge management, aiming to address information overload on intranets and the web by enabling semantic rather than keyword-based access. Central to the framework is the Ontology Inference Layer (OIL), which integrates description logics for formal reasoning, frame-based modeling for intuitive structuring, and web standards (XML/RDF) for interoperability, balancing expressivity with decidability. The tool environment supports ontology-driven querying, RDF/XML repositories, annotation technologies, and visualization tools, while industrial case studies in insurance (Swiss Life), telecom (BT call centres), and energy (EnerSearch) demonstrate practical applications of semantic access to heterogeneous data. Methodologically, the project builds on approaches like CommonKADS to formalize ontology engineering and business case development. For Knowledge Representation and Reasoning, the paper is significant because it shows how formal logic-based ontologies can be applied to real-world problems, laying the groundwork for the Semantic Web and evolving into standards like OWL that underpin today’s knowledge graphs and AI reasoning systems.
The knowledge-based economy and the triple helix model This paper by Leydesdorff (2012) introduced me to the meaning and the challenges of the knowledge-based economy (KBE), which was challenging for me to understand, given that I do not have a background in the topic. As per my understanding, KBE is an economy that is driven by innovation that is in turn based on knowledge. However, the paper itself notes that the concept of KBE is not well defined, and lacks of the dynamic interplay in such an economy. That is why the paper proposes the Triple Helix model, which is a neo-evolutionary analytical framework designed to define, model and explain the complex, non-linear dynamics of the KBE. The main propositions of the model include that that there are three main pillars (helices) of innovation; University, Industry and Government. They respectively play their roles in novelty generation, wealth generation, and govenrance and control. The non-linear interplay between these sub-dynamics is what dialectically constructs innovation. The paper also introduced me to the concept of codified knowledge, which is knowledge that is abstracted in a system of symbols, and therefore is decontextualised and can be communicated and shared among different domains and institutions, or 'traded in a market,' as the paper puts it. I see codified knowledge as the most relevant part to the module topic.

Artefacts

Formative Activity: Questions on Readings

The formative activity of this unit included answering questions on the three reading materials. Below are my answers:

Activity 1: Questions on Bimba et al. (2016)

Q1: What are the 3 knowledge bases identified by the authors?

The three knowledge bases identified by the authors are: Linguistic Knowledge Base, Expert Knowledge Base and Cognitive Knowledge Base.

Q2: What are the differences between them?

Linguistic Knowledge Base focuses on the representation of knowledge through the relationsips between words, meanings and concepts in natural language. Expert Knowledge Base focuses on capturing and presenting knowledge in the form of logical rules or fuzzy rules. Cognitive Knowledge Base aims to model human cognitive processes and structures to represent knowledge in a way that mimics human thinking and reasoning.

Q3: How are these related to ontologies?

Ontogies are presented in the paper as a separate knowledge model, but a conenction can be made, seeing that ontology is the underlying foundation for the three bases, as ontology -in this context- means the representation of knowledge as taxonomy of concepts, which can be applied through linguistic, expert or cognitive knowledge bases.

Activity 2: Questions on Leydesdorff (2012)

Q1: How does the author define a ‘Knowledge based economy’?

The paper declares that there are hardships in defining the concept of knowledge-based economy (KBE), but generally, it can be defined as an economy that is driven by innovation that is in turn based on knowledge.

Q2: What is the triple helix model?

The triple helix model is a neo-evolutionary analytical framework designed to define, model and explain the complex, non-linear dynamics of the KBE. The main propositions of the model include that that there are three main pillars (helices) of innovation; University, Industry and Government. They respectively play their roles in novelty generation, wealth generation, and govenrance and control.

Activity 3: Questions on Fensel et al. (2000)

Q1: What are the components of the Onto Knowledge model/ framework?

There are six components of the Onto Knowledge framework, which are: 1) Ontology-based user interaction, built in Ontology Inference Layer (OIL); 2) Ontology construction tools; 3) Data repository layer (RDF/XML repository); 4) Query engine (XML or RDF query engines); 5) Annotation and extraction tools; 6)Visualisation tools.

Q2: What are the example use cases given for the framework?

There are three example use cases given for the framework, which are: 1) Swiss Life insurance company, to enable semantic access to their intranet data; 2) British Telecom (BT) call centres, to improve customer support through semantic access to heterogeneous data; 3) EnerSearch, an energy company, to facilitate semantic access to energy-related information.

Artefacts

Mind Map: Knowledge Models

The following mind map summarises the different knowledge models presented in Bimba et al. (2016):

Figure 1: Mind Map of Knowledge Models based on Bimba et al. (2016)

Reference List

Bimba, A.T. et al. (2016) ‘Towards Knowledge Modeling and Manipulation technologies: a Survey’, International Journal of Information Management, 36(6), pp. 857–871. Available at: https://doi.org/10.1016/j.ijinfomgt.2016.05.022.

Fensel, D. et al. (2000) ‘“On-To-Knowledge: Ontology-based Tools for Knowledge Management”’, in eBusiness and eWork.

Leydesdorff, L. (2010) ‘The knowledge-based Economy and the Triple Helix Model’, Annual Review of Information Science and Technology, 44(1), pp. 365–417. Available at: https://doi.org/10.1002/aris.2010.1440440116.