Ontology, Security, and NLP

1
Ontology, Security, and NLP

• -- A discussion for the Security Reading Group of
  CSE_at_USC
• -- Xinliang Zheng

2
Outline

• Ontological needs in information security
• Incorporates natural language data to information
  security
• Ontology semantics in NLP
• NLP with ontology applied in information security
  ?

3
Ontological needs in information security

• Ontological approach is a valuable resource in
  systematizing the phenomena in the purview,
  enabling the modular approach, and predicting new
  phenomena in the security domain.
• Phenomena types of attack and countermeasure.

4
Needs a common language

• A set of security-related terms with defined
  meanings
• A taxonomy of the terms
• Showing the relations between those terms

5
The intension of the common language

• Talk more understandably with others about
  security
• Gather, organize, and record security
  information
• Extract data from security information
• Sharing of security information
• Use the information to evaluate and decide on
  proper courses of action
• Use the information to determine effects of
  actions over time.

6
Incorporates natural language data to information
security

• Why
• Sysadmin logs are written in a sublanguage of a
  natural language
• Information hiding (steganography, NL
  watermarking) depends on NLP
• Downgrading will provide automatic filtering of
  sensitive information from documents intended for
  dissemination
• Documents in natural language can be scanned for
  detecting possible intellectual property leakage.

7
Incorporates natural language data to information
security

• What
• Inclusion of natural language data sources as an
  integral part of the overall data sources in
  information security applications
• Formal specification of the information security
  community know-how for the support of routine and
  time-efficient measures to prevent and counteract
  security attacks.

8
Ontology semantics in NLP

• Natural Language Processing (NLP) can be seen, in
  security terms, as the effort of decoding the
  meaning of a text from its surface form.
• Ontological semantics is an approach for
  extracting and representing meaning of natural
  language texts as well as synthesizing natural
  language texts based on representations of their
  meaning.

9
Architecture of prototypical ontological
semantics for NLP

• A set of static knowledge sources
• Knowledge representation languages for specifying
  meaning strutures, ontologies and lexicons
• A set of processing modules, namely, a semantic
  analyzer and a semantic text generator.

10
NLP with ontology applied in information security
?

• How to incorporates them together?
• Current status
• Purdue, NWS, Bell Lab is doing some research on
  this.
• Reference Papers
• http//delivery.acm.org/10.1145/370000/366190/p51-
  tallah.pdf?key1366190key28392648701collACMdl
  ACMCFID18466038CFTOKEN83431105
• http//delivery.acm.org/10.1145/510000/508180/p53
  -raskin.pdf?key1508180key29003648701collACMd
  lACMCFID18466038CFTOKEN83431105