Semantics Demystified

Home » Semantics Demystified

Report by ISITC Europe to demystify Semantics.

Author: Martin, Sexton, Principal Consultant at London Market Systems, with contributions from Richard Beatch at Bloomberg

With the potential impact of BCBS 239 and financial institutions focus on governance and ownership of risk, there is an appreciation that key to supporting governance is the deployment of the appropriate IT and data architectures. That raises the question, “Does semantics have a role to play in supporting governance?”

Before examining this question, it is appropriate to provide some context to semantics itself.

So, what is semantics?

Put simply it is the study of meaning and without semantics there is no context for understanding information.

A key initiative in the financial sector is the Financial Industry Business Ontology (or FIBO), try searching for FIBO on the Internet and what do you get? Lots of videos of a fitness and wellbeing fair. This experiment highlights the need for being able to understand the context of information.

Semantic web technologies can be used to represent two main types of models:

♦   A Taxonomy is a hierarchy of concepts, the relations are normally parent/child, superClass/subClass, or broader/narrower.

♦   An Ontology is a formal model of concepts, the relationships that connect them, and constraints on the ways that concepts and relationships can be combined. These relationships can be complex and may not result in a single hierarchical structure. An Ontology can use a taxonomy or a set of taxonomies as its base.

Given the vast number of terms, abbreviations and acronyms used in the world of semantics, the following Glossary has been provided to bring the reader up to speed on this subject.

URIUniform Resource Identifier is a unique string of characters to identify a concept and infers an ASCII encoding. Also see IRI.

Concept Is a unit thought, idea or meaning. A Concept in the semantic domain is normally associated with a unique identifier; this is called a IRI (or URI).
Dublin Core Is an RDF vocabulary that supports key metadata terms, this includes the ability to identify the creator, support versioning, among others.  This vocabulary is used to annotate ontologies constructed using OWL or SKOS. Dublin Core is also an ISO standard, ISO 15836.
IRI Internationalized Resource Identifier is a unique string of characters to identify a concept, in a Unicode encoding, sometimes referred to as URI. However URI infers an ASCII encoding.
JSON-LD JavaScript Object Notation for Linked Data, is a method of encoding Linked Data. For further information, please refer to: https://en.wikipedia.org/wiki/JSON-LD
Ontology A formal model of concepts, the relationships that connect them, and constraints on the ways that concepts and relationships can be combined. These are complex relationships and may not result in a simple hierarchical structure.
OWL The W3C Web Ontology Language is a Semantic Web language designed to represent rich and complex knowledge about things, groups of things, and relations between things. It allows one to represent hierarchical class relationships and capture properties, constraints among other things. For further information, please refer to: https://www.w3.org/OWL/. There are various syntax conventions by which OWL can be represented (see “TTL” below).
Provenance Is the history of the ownership and dissemination of objects.
PROV-O Prov-O is a W3C initiate to capture It provides a set of classes, properties, and restrictions that can be used to represent and interchange provenance.
RDF/XML Resource Description Framework, normally encoded in XML, provides the framework to define the relationship between concepts; this is usually referred to as a “Triple”.  For further information, please refer to: https://www.w3.org/TR/REC-rdf-syntax/
SKOS Simple Knowledge Organization System is a W3C recommendation designed for representing classification schemes and taxonomies. Like OWL, SKOS is an RDF-based vocabulary. Unlike the class hierarchy one might develop in OWL, SKOS provides the ability to create hierarchies that utilize different types of relationships, e.g., is-a-part/member-of and as such, it provides the opportunity to support classifications and taxonomies across a broad range of information and use cases. For further information, please refer to: https://www.w3.org/SKOS/
Taxonomy Is a hierarchy of concepts, the relations are normally parent/child, superclass/subClass, or broader/narrower.
Triple The semantic web at the atomic level can be broken down into Triples. An OWL or SKOS based model can also be expressed in Triples.  A Triple provides the capability to support data linkages between concepts and comprises three of parts:

Subject The thing upon which the relationship/modification is to be applied to.
Predicate Describes the relationship between the Subject and Object.
Object The thing the Predicate is acting upon, also known as the target.

The object of one Triple can be the subject of another triple, allowing one to represent relationships between items, this lends itself being visualising data in a graphical form and also providing linkages between model objects.

Subject Predicate Object
Equities Can take the form of Common/ordinary shares (S)
Debt Instrument pays a Coupon

Subjects and Objects are Concepts and in ontologies are uniquely identified using IRIs.

TTL Terse RDF Triple Language or Turtle/TTL is a syntax convention that represents the Web Ontology Language (OWL). For further information on this OWL syntax and details regarding how it is structured please refer to https://www.w3.org/TeamSubmission/turtle/.

There are a number of semantic web based initiatives within the financial sector. The key ones are summarised in the following paragraphs.

FIBO

The Financial Industry Business Ontology (or FIBO), initiated by Michael Bennett and developed through the EDM Council, comprises a number of financial ontologies that provide meaning to legal/business entities and financial products. For further information please refer to: https://spec.edmcouncil.org/fibo

ISO 20022 Semantic Proof of Concept

There is an initiative underway by ISO Technical Committee 68 SC9 Working Group 1 (WG1) to improve and enrich the ISO 20022 standard and to that end the working group is developing a proof of concept portal to show equivalences between the main industry standards, FpML, FIX and ISO 20022 message models. Each standard is being captured within an OWL based ontology.

ISO CFI Semantic Model

ISO TC68 SC8 Working Group 6 (WG6) has created a work stream to investigate the appropriate electronic representation of ISO 10962 (CFI) using semantic web technologies. Taking the form of a service, the intension is to allow consumers to download the standard in various formats and integrate this information automatically with enterprise applications. Currently, the official version of the ISO standard is only available as a PDF document via the ISO website.

A semantic based ontology has been proposed and is visualised below.

semanTYCS-DEMYSTIFIED model
Figure 1. Visual Representation of a snippet of the CFI ontology (Author: Richard Beatch – Bloomberg)

 

 

 

How can semantic web technology benefit the enterprise?

To start, data objects and terms can be captured as Concepts within the organisation’s semantic based common data dictionary. When one adds relationships to Concepts, the Dictionary’s value becomes enhanced. These data lineages will enable the capability to support valuations, risk reference data and trade lifecycle processes.

Ontologies such as FIBO can form the basis of any initiative and by adding vocabularies, such as Dublin Core and Prov-O to support traceability and provenance, one can construct a service that can become core to the organisation’s governance and thus enable adherence to BCBS 239.

To attain corporate fitness and wellbeing, there is a growing belief that semantic web technologies does have a role to play.

+ Introduction

Report by ISITC Europe to demystify Semantics.

Author: Martin, Sexton, Principal Consultant at London Market Systems, with contributions from Richard Beatch at Bloomberg

With the potential impact of BCBS 239 and financial institutions focus on governance and ownership of risk, there is an appreciation that key to supporting governance is the deployment of the appropriate IT and data architectures. That raises the question, “Does semantics have a role to play in supporting governance?”

Before examining this question, it is appropriate to provide some context to semantics itself.

So, what is semantics?

Put simply it is the study of meaning and without semantics there is no context for understanding information.

A key initiative in the financial sector is the Financial Industry Business Ontology (or FIBO), try searching for FIBO on the Internet and what do you get? Lots of videos of a fitness and wellbeing fair. This experiment highlights the need for being able to understand the context of information.

Semantic web technologies can be used to represent two main types of models:

♦   A Taxonomy is a hierarchy of concepts, the relations are normally parent/child, superClass/subClass, or broader/narrower.

♦   An Ontology is a formal model of concepts, the relationships that connect them, and constraints on the ways that concepts and relationships can be combined. These relationships can be complex and may not result in a single hierarchical structure. An Ontology can use a taxonomy or a set of taxonomies as its base.

+ Glossary

Given the vast number of terms, abbreviations and acronyms used in the world of semantics, the following Glossary has been provided to bring the reader up to speed on this subject.

URIUniform Resource Identifier is a unique string of characters to identify a concept and infers an ASCII encoding. Also see IRI.

Concept Is a unit thought, idea or meaning. A Concept in the semantic domain is normally associated with a unique identifier; this is called a IRI (or URI).
Dublin Core Is an RDF vocabulary that supports key metadata terms, this includes the ability to identify the creator, support versioning, among others.  This vocabulary is used to annotate ontologies constructed using OWL or SKOS. Dublin Core is also an ISO standard, ISO 15836.
IRI Internationalized Resource Identifier is a unique string of characters to identify a concept, in a Unicode encoding, sometimes referred to as URI. However URI infers an ASCII encoding.
JSON-LD JavaScript Object Notation for Linked Data, is a method of encoding Linked Data. For further information, please refer to: https://en.wikipedia.org/wiki/JSON-LD
Ontology A formal model of concepts, the relationships that connect them, and constraints on the ways that concepts and relationships can be combined. These are complex relationships and may not result in a simple hierarchical structure.
OWL The W3C Web Ontology Language is a Semantic Web language designed to represent rich and complex knowledge about things, groups of things, and relations between things. It allows one to represent hierarchical class relationships and capture properties, constraints among other things. For further information, please refer to: https://www.w3.org/OWL/. There are various syntax conventions by which OWL can be represented (see “TTL” below).
Provenance Is the history of the ownership and dissemination of objects.
PROV-O Prov-O is a W3C initiate to capture It provides a set of classes, properties, and restrictions that can be used to represent and interchange provenance.
RDF/XML Resource Description Framework, normally encoded in XML, provides the framework to define the relationship between concepts; this is usually referred to as a “Triple”.  For further information, please refer to: https://www.w3.org/TR/REC-rdf-syntax/
SKOS Simple Knowledge Organization System is a W3C recommendation designed for representing classification schemes and taxonomies. Like OWL, SKOS is an RDF-based vocabulary. Unlike the class hierarchy one might develop in OWL, SKOS provides the ability to create hierarchies that utilize different types of relationships, e.g., is-a-part/member-of and as such, it provides the opportunity to support classifications and taxonomies across a broad range of information and use cases. For further information, please refer to: https://www.w3.org/SKOS/
Taxonomy Is a hierarchy of concepts, the relations are normally parent/child, superclass/subClass, or broader/narrower.
Triple The semantic web at the atomic level can be broken down into Triples. An OWL or SKOS based model can also be expressed in Triples.  A Triple provides the capability to support data linkages between concepts and comprises three of parts:

Subject The thing upon which the relationship/modification is to be applied to.
Predicate Describes the relationship between the Subject and Object.
Object The thing the Predicate is acting upon, also known as the target.

The object of one Triple can be the subject of another triple, allowing one to represent relationships between items, this lends itself being visualising data in a graphical form and also providing linkages between model objects.

Subject Predicate Object
Equities Can take the form of Common/ordinary shares (S)
Debt Instrument pays a Coupon

Subjects and Objects are Concepts and in ontologies are uniquely identified using IRIs.

TTL Terse RDF Triple Language or Turtle/TTL is a syntax convention that represents the Web Ontology Language (OWL). For further information on this OWL syntax and details regarding how it is structured please refer to https://www.w3.org/TeamSubmission/turtle/.
+ Semantic Web based Initiatives

There are a number of semantic web based initiatives within the financial sector. The key ones are summarised in the following paragraphs.

FIBO

The Financial Industry Business Ontology (or FIBO), initiated by Michael Bennett and developed through the EDM Council, comprises a number of financial ontologies that provide meaning to legal/business entities and financial products. For further information please refer to: https://spec.edmcouncil.org/fibo

ISO 20022 Semantic Proof of Concept

There is an initiative underway by ISO Technical Committee 68 SC9 Working Group 1 (WG1) to improve and enrich the ISO 20022 standard and to that end the working group is developing a proof of concept portal to show equivalences between the main industry standards, FpML, FIX and ISO 20022 message models. Each standard is being captured within an OWL based ontology.

ISO CFI Semantic Model

ISO TC68 SC8 Working Group 6 (WG6) has created a work stream to investigate the appropriate electronic representation of ISO 10962 (CFI) using semantic web technologies. Taking the form of a service, the intension is to allow consumers to download the standard in various formats and integrate this information automatically with enterprise applications. Currently, the official version of the ISO standard is only available as a PDF document via the ISO website.

A semantic based ontology has been proposed and is visualised below.

semanTYCS-DEMYSTIFIED model
Figure 1. Visual Representation of a snippet of the CFI ontology (Author: Richard Beatch – Bloomberg)

 

 

 

+ Final Thoughts

How can semantic web technology benefit the enterprise?

To start, data objects and terms can be captured as Concepts within the organisation’s semantic based common data dictionary. When one adds relationships to Concepts, the Dictionary’s value becomes enhanced. These data lineages will enable the capability to support valuations, risk reference data and trade lifecycle processes.

Ontologies such as FIBO can form the basis of any initiative and by adding vocabularies, such as Dublin Core and Prov-O to support traceability and provenance, one can construct a service that can become core to the organisation’s governance and thus enable adherence to BCBS 239.

To attain corporate fitness and wellbeing, there is a growing belief that semantic web technologies does have a role to play.

Author: Martin, Sexton, Principal Consultant at London Market Systems, with contributions from Richard Beatch at Bloomberg