Today's experiment with ChatGPT4 was seeing how well it could support a creating a SKOS taxonomy from text I copy and pasted out of one of my Cybersecurity Science blogs, "A Standardized Vocabulary for Evaluating the Impact of Cyber Defense Decisions on Adversary Behavior" which contains an effects vocabulary from NIST 800-160 Vol 2 Rev 1 dated December 2021.
SKOS (Simple Knowledge Organization System) is a widely-used standard for representing knowledge organization systems, such as taxonomies, thesauri, and classification schemes. Here are some of the benefits of using SKOS:
- Interoperability: SKOS is a widely recognized and adopted standard, making it easy to share and exchange knowledge organization systems across different applications and platforms. This can improve interoperability and facilitate data integration and reuse.
- Consistency: SKOS provides a consistent way to represent knowledge organization systems, making it easier to manage and maintain them. This can help ensure that the terminology used is consistent across different applications and systems, improving data quality and reducing errors.
- Flexibility: SKOS is a flexible standard that can be used to represent a wide range of knowledge organization systems, from simple taxonomies to complex ontologies. This makes it suitable for a wide range of applications, from library and information science to data-driven industries like finance and healthcare.
- Reusability: SKOS enables the reuse of existing knowledge organization systems, making it easier to build on existing work rather than starting from scratch. This can save time and resources, and improve the accuracy and consistency of knowledge organization.
- Machine-readability: SKOS is designed to be machine-readable, making it easier to process and analyze knowledge organization systems using software tools. This can enable advanced applications like automated indexing, semantic search, and machine learning.
The input text to ChatGPT4 was section III B & C which contained text describing the vocabulary. I crafted a prompt to create a SKOS taxonomy from the text that included the 5 high-level effects and the narrower effects that support them, the 14 specific effects, the impact on risk of all effects, and details about the original NIST source document. The following is the ChatGPT4 response.
@prefix ex: <http://example.org/effects/> .
@prefix skos: <http://www.w3.org/2004/02/skos/core#> .
ex:EffectsVocabulary a skos:ConceptScheme ;
skos:prefLabel "Effects Vocabulary from NIST SP 800-160 Vol 2 Rev 1 (December 2021)" ;
skos:definition "A standardized vocabulary of high-level and specific effects on adversaries,
based on NIST SP 800-160 Vol 2 Rev 1 (December 2021). It is designed for adversary-oriented analysis
and includes impact on risk information for each effect." .
ex:Redirect a skos:Concept ;
skos:prefLabel "Redirect" ;
skos:definition "Divert the adversary's activities to a less harmful or less critical target." ;
skos:narrower ex:Deter, ex:Divert, ex:Deceive ;
skos:inScheme ex:EffectsVocabulary ;
skos:note "Impact on Risk: Decreases likelihood" .
ex:Preclude a skos:Concept ;
skos:prefLabel "Preclude" ;
skos:definition "Prevent the adversary from conducting a successful attack." ;
skos:narrower ex:Expunge, ex:Preempt, ex:Negate ;
skos:inScheme ex:EffectsVocabulary ;
skos:note "Impact on Risk: Decreases likelihood" .
ex:Impede a skos:Concept ;
skos:prefLabel "Impede" ;
skos:definition "Hinder the adversary's ability to conduct an attack." ;
skos:narrower ex:Contain, ex:Degrade, ex:Delay, ex:Exert ;
skos:inScheme ex:EffectsVocabulary ;
skos:note "Impact on Risk: Decreases likelihood" .
ex:Limit a skos:Concept ;
skos:prefLabel "Limit" ;
skos:definition "Reduce the impact of a successful attack." ;
skos:narrower ex:Shorten, ex:Reduce ;
skos:inScheme ex:EffectsVocabulary ;
skos:note "Impact on Risk: Decreases impact" .
ex:Expose a skos:Concept ;
skos:prefLabel "Expose" ;
skos:definition "Reveal the adversary's activities to better prepare defenders for future attacks." ;
skos:narrower ex:Detect, ex:Reveal, ex:Scrutinize ;
skos:inScheme ex:EffectsVocabulary ;
skos:note "Impact on Risk: Decreases likelihood and impact" .
# The specific effects concepts are defined here, with their relationships to the high-level effects concepts
ex:Deter a skos:Concept ;
skos:prefLabel "Deter" ;
skos:definition "Discourage adversaries from attempting an attack through threats or other means of dissuasion." ;
skos:broader ex:Redirect ;
skos:inScheme ex:EffectsVocabulary ;
skos:note "Impact on Risk: Decreases likelihood" .
ex:Divert a skos:Concept ;
skos:prefLabel "Divert" ;
skos:definition "Channel adversaries away from high-value targets by creating false targets or deception." ;
skos:broader ex:Redirect ;
skos:inScheme ex:EffectsVocabulary ;
skos:note "Impact on Risk: Decreases likelihood" .
ex:Deceive a skos:Concept ;
skos:prefLabel "Deceive" ;
skos:definition "Mislead adversaries into taking ineffective or suboptimal actions." ;
skos:broader ex:Redirect ;
skos:inScheme ex:EffectsVocabulary ;
skos:note "Impact on Risk: Decreases likelihood" .
ex:Expunge a skos:Concept ;
skos:prefLabel "Expunge" ;
skos:definition "Eradicate adversaries from the system or network entirely." ;
skos:broader ex:Preclude ;
skos:inScheme ex:EffectsVocabulary ;
skos:note "Impact on Risk: Decreases likelihood" .
ex:Preempt a skos:Concept ;
skos:prefLabel "Preempt" ;
skos:definition "Actively counter adversary activities before they can cause harm." ;
skos:broader ex:Preclude ;
skos:inScheme ex:EffectsVocabulary ;
skos:note "Impact on Risk: Decreases likelihood" .
ex:Negate a skos:Concept ;
skos:prefLabel "Negate" ;
skos:definition "Neutralize the adversary's ability to exploit a vulnerability or conduct an attack." ;
skos:broader ex:Preclude ;
skos:inScheme ex:EffectsVocabulary ;
skos:note "Impact on Risk: Decreases likelihood" .
ex:Contain a skos:Concept ;
skos:prefLabel "Contain" ;
skos:definition "Restrict the adversary's movement or access within a system or network." ;
skos:broader ex:Impede ;
skos:inScheme ex:EffectsVocabulary ;
skos:note "Impact on Risk: Decreases likelihood" .
ex:Degrade a skos:Concept ;
skos:prefLabel "Degrade" ;
skos:definition "Reduce the adversary's capabilities or the effectiveness of their tools and techniques." ;
skos:broader ex:Impede ;
skos:inScheme ex:EffectsVocabulary ;
skos:note "Impact on Risk: Decreases likelihood" .
ex:Delay a skos:Concept ;
skos:prefLabel "Delay" ;
skos:definition "Slow down the adversary's progress or actions, buying time for defenders to respond." ;
skos:broader ex:Impede ;
skos:inScheme ex:EffectsVocabulary ;
skos:note "Impact on Risk: Decreases likelihood" .
ex:Exert a skos:Concept ;
skos:prefLabel "Exert" ;
skos:definition "Apply pressure on the adversary, forcing them to expend more resources or take less desirable actions." ;
skos:broader ex:Impede ;
skos:inScheme ex:EffectsVocabulary ;
skos:note "Impact on Risk: Decreases likelihood" .
ex:Shorten a skos:Concept ;
skos:prefLabel "Shorten" ;
skos:definition "Reduce the duration of a successful attack, limiting the adversary's ability to cause damage." ;
skos:broader ex:Limit ;
skos:inScheme ex:EffectsVocabulary ;
skos:note "Impact on Risk: Decreases impact" .
ex:Reduce a skos:Concept ;
skos:prefLabel "Reduce" ;
skos:definition "Diminish the extent or severity of the
consequences resulting from a successful attack." ;
skos:broader ex:Limit ;
skos:inScheme ex:EffectsVocabulary ;
skos:note "Impact on Risk: Decreases impact" .
ex:Detect a skos:Concept ;
skos:prefLabel "Detect" ;
skos:definition "Identify adversary activities, enabling defenders to respond more effectively." ;
skos:broader ex:Expose ;
skos:inScheme ex:EffectsVocabulary ;
skos:note "Impact on Risk: Decreases likelihood and impact" .
ex:Reveal a skos:Concept ;
skos:prefLabel "Reveal" ;
skos:definition "Uncover information about the adversary's tactics, techniques, and procedures." ;
skos:broader ex:Expose ;
skos:inScheme ex:EffectsVocabulary ;
skos:note "Impact on Risk: Decreases likelihood and impact" .
ex:Scrutinize a skos:Concept ;
skos:prefLabel "Scrutinize" ;
skos:definition "Examine and analyze the adversary's activities in detail to improve defensive capabilities." ;
skos:broader ex:Expose ;
skos:inScheme ex:EffectsVocabulary ;
skos:note "Impact on Risk: Decreases likelihood and impact" .
I then asked it if it could use the SKOS taxonomy to generate a network diagram in Mermaid.js notation.
graph LR
subgraph Effects
Redirect[Redirect]
Preclude[Preclude]
Impede[Impede]
Limit[Limit]
Expose[Expose]
end
subgraph Redirect
Deter[Deter]
Divert[Divert]
Deceive[Deceive]
end
subgraph Preclude
Expunge[Expunge]
Preempt[Preempt]
Negate[Negate]
end
subgraph Impede
Contain[Contain]
Degrade[Degrade]
Delay[Delay]
Exert[Exert]
end
subgraph Limit
Shorten[Shorten]
Reduce[Reduce]
end
subgraph Expose
Detect[Detect]
Reveal[Reveal]
Scrutinize[Scrutinize]
end
Redirect --> Deter
Redirect --> Divert
Redirect --> Deceive
Preclude --> Expunge
Preclude --> Preempt
Preclude --> Negate
Impede --> Contain
Impede --> Degrade
Impede --> Delay
Impede --> Exert
Limit --> Shorten
Limit --> Reduce
Expose --> Detect
Expose --> Reveal
Expose --> Scrutinize
To generate a diagram from this notation, you can copy and paste this code into the Mermaid Live Editor (https://mermaid-js.github.io/mermaid-live-editor/). This will produce the following diagram:
I have to say I'm pretty impressed with ChatGPT's ability to generate the above just from the text contained in my blog post about the effects vocabulary. I'm excited for what this means for domain experts who might not need to enlist the expertise of a Knowledge Engineers or Ontologists to quickly capture domain specific knowledge.