Semantic annotation for knowledge explicitation in a product lifecycle management context

TL;DR: This work collects a number of literature that applied semantic annotations on different objects, and classify them according to the subject being described in an enterprise architecture framework, and identifies the existing drawbacks.

About: This article is published in Computers in Industry.The article was published on 2015-08-01 and is currently open access. It has received 46 citations till now. The article focuses on the topics: Semantic interoperability & Semantic computing.

Summary

1. Introduction

• In manufacturing enterprises, the Product Lifecycle Management (PLM) approach has been considered as an essential solution for improving the product competitive ability.
• This results in a kind of “tower of Babel”, where each application is considered as an island in the middle of the ocean of information, managed by stakeholders along the life cycle of a product.
• The mutual understanding of the semantics that is embedded inside the exchanged information is the cornerstone in the quest for semantic interoperability.
• The rest of this paper is organized as follows: Section 2 presents the definitions of annotation and semantic annotation.
• Section 5 concludes this paper and highlights their on-going and further research works.

2. Annotation and Semantic Annotation

• It has special usages in different contexts.
• In the mechanical drawing, an annotation is a snippet of text or symbol with specific meanings that illustrates the corresponding annotated part.
• In order to distinguish the semantic annotation from the other annotations, several kinds of classifications are proposed.
• Lin [9] considered it as “an approach to link ontologies to the original information sources”.
• The domain semantics, which describes the context and the meaning of an annotated element in a specific domain; the structure semantics, which describes the interrelations between an annotated element and the other elements related to it.

3. The Investigation of Semantic Annotation Researches

• In the last decade, several surveys of semantic annotation researches have already been made with different focuses.
• Hanbury [17] summarized five types of image annotation methods and then used it to analyse ten annotated image datasets.
• A number of self-defined requirements are used as the basis to compare the semantic annotation approaches in the surveys [14] and [16].
• The survey [16] is the only one that concerned the annotation on a specific kind of model (e.g. workflows).
• Little attention has been paid to the in-depth study and comparison of the methods, especially from the formalization perspective.

3.1 The Illustration of Semantic Annotation Researches

• With the supports of the ontologies from multiple domains and different levels, the semantic annotation approaches are widely studied and applied in diverse contexts.
• Moreover, for each chosen subject, the authors will present the analysis of two or three research works that applied semantic annotation on it as examples.
• It provides for each extracted NE with two kinds of links: one link to the most specific class in KIM ontology to specify the named-entity type and the other link to the specific individual in knowledge base.

3.2 The Comparison of Semantic Annotation Researches

• From the General Point of View As it is illustrated in Table 1, an overall comparison of the above-mentioned examples is presented, also known as 2.1 The Comparison.
• The annotations are used to attach machine-readable semantics to the annotated elements and to obtain semantic reasoning supports (e.g. inference).
• (4) Element type ④, which contains the relations between the annotated element (that identified by ①) and its domain or structure semantics (② or ③).
• To the contrary, for those researches which embed references in the annotation objects, their schemas, such as (a) and (i), do not contain this type of elements.
• In (c), relations are represented as three types of assertions.

4 Existing Drawbacks and Possible Research Directions

• As discussed in previous sections, the authors found that despite lots of efforts have been made in semantic annotation researches, at least, three existing drawbacks can be noted.
• The formalization of semantic annotations is not the focus in some of above- mentioned researches ([29], [30], [35], [36] and [41]), where it is only considered as a kind of “is a” association between an annotated element and an ontology concept.
• These schemas are difficult to be reused in other researches but the studied ones.
• In the research [22], it is used to express modelling construct and support models transformations.
• The research [23] is the only one that proposed to combine both aspects of semantics in the reasoning phase.

5. Conclusion

• In a PLM environment, various kinds of representations are used to capture and describe the knowledge related to a product along its life cycle.
• During the collaboration, a mutual understanding of the semantics inside these shared and exchanged knowledge representations is the foundation to achieve the semantic interoperability.
• The authors present a survey on a number of collected semantic annotation literature and provide a detailed comparison and discussion, especially from the formalization perspective.
• Based on this survey, several existing drawbacks and possible research directions are identified.

Figures

Figure 5. The Semantic Annotation Schema from Lin (2008) [9]

Figure 9. The Semantic Annotation Schema from Talantikite, et al. (2009) [8] Patil et al. [41] proposed the MWSAF, a framework for semi-automatically

Figure 11. The Comparison of Semantic Annotation Schemas

Figure 4. The Semantic Annotation Schema from Di Francescomarino (2011) [26] Lin [9] proposed a semantic annotation framework to support the discovery and

Figure 3. The Semantic Annotation Schema from Liao (2013) [23]

Figure 8. The Semantic Annotation Schema from Ma et al. (2011) [37]

Table 1. The Comparison of Semantic Annotation Researches

Figure 7. The Semantic Annotation Schema from Li (2012) [34]

Figure 1. The Classification of the Collected Literature based on the Zachman Framework

Figure 2. The Semantic Annotation Schema from Boudjlida et al. (2007) [22] Liao [23] studied the issues of the interoperability, especially the semantic

Figure 10. The Semantic Annotation Schema from Kopecký et al. (2007) [40]

Frequently asked questions

Q1. What are the contributions in "Semantic annotation for knowledge explicitation in a product lifecycle management context: a survey" ?

The authors collect a number of literature that applied semantic annotations on different objects, and classify them according to the subject being described in an enterprise architecture framework. In this paper, a detailed survey, especially from the formalization perspective, is presented to identify the existing drawbacks and to point out the possible research directions. 

Q2. What are the future works in "Semantic annotation for knowledge explicitation in a product lifecycle management context: a survey" ?

In their on-going and future research works, the authors intend to enrich this survey with a more complete analysis from different perspectives, such as implementation and performance perspectives. 

Q3. What is the cornerstone in the quest for semantic interoperability?

The mutual understanding of the semantics that is embedded inside the exchanged information is the cornerstone in the quest for semantic interoperability. 

Q4. What is the only concern in some of above-mentioned researches?

Making explicit the domain semantics is the only concern in some of above-mentioned researches ([8], [29], [30], [34], [35], [36] and [41]), where the structure semantics is ignored. 

Q5. What is the challenge to carry out future researches in maintaining the consistency of semantic annotations?

Along with the versioning of annotatedobjects and the evolution of ontologies, there remains a promising challenge to carry out future researches in maintaining the consistency of semantic annotations. 

Q6. What is the purpose of adding semantic annotations to a web service?

In the Detailed Representation Perspective, adding semantic annotations to a web service is mainly for supporting the automatic verification of certain tasks, which must be executed before or during invocation of corresponding services [40]. 

Q7. What are the main issues that were not addressed by the surveys?

The authors can find that the surveys [13], [14], [15], [19] and [20] were mainly focusingon documents, as well as the surveys [17] and [18] paid major attention to images or videos. 

Q8. What is the purpose of this paper?

The objective of this paper is to present a detailed survey of the collected semanticannotation literature, especially from the formalization perspective. 

Q9. What is the main focus of the researches?

The formalization of semantic annotations is not the focus in some of above-mentioned researches ([29], [30], [35], [36] and [41]), where it is only considered as a kind of “is a” association between an annotated element and an ontology concept. 

Q10. What is the semantic annotation in a business process diagram?

On one hand, a semantic annotation is represented as the “textual annotations” in a business process diagram by using a “@” symbol with the name of the selected ontology class.