fake tab top


Domain Specific Ontologies

The idea of “Domain Specific Ontologies” is fundamental.  Take, for example the following use cases:

  • The field of Artificial Intelligence often deals with translation between “Domains of Discourse.”
  • In the field of Distributed Simulation we often create “Federation Object Models” to define the interchange of data between simulations to accomplish federation or exercise goals.

  • A Data Base Architect (DBA) designs database schema and stored procedures to store enterprise relevant data, and efficiently search and extract it in forms relevant to diverse enterprise needs.

In all these fields, the concept of Domain Specific Ontologies is fundamental.

We present an illustrative example of how differing purpose creates appropriately varying Domains of Discourse:

The Automobile wheel example:


How many wheels does the car have?


  1. Buyer, sitting in the car: The Car has one wheel, the steering wheel

  2. Salesperson standing outside the car:  The car has 4 wheels, and they are 20” alloys.

  3. The corporate Tire Warranty Manager: The car has 5 wheels with 5 tires, which can vary on size and model based on the wheel size,.

  4. The Automotive Engine Engineer, I suppose the car has about 156 wheels, plus the 5 road wheels, steering wheel, and not including the sound system.

The key issue here is that for each of their purposes, each answer is correct.

Doman Specific Ontologies for High Accuracy Translation Triage (HATT)

ArtisTech developed an approach to provide high-accuracy document search and triage in a target Foreign Language (FL) using a small domain specific ontology which includes highly accurate translated terms and phrases.  The prototype of the HATT approach is called OntoFLex.  OntoFLex allows the user to build an acyclic graph with relational links including both domain and metadata contents.  This tool then uses general and language-specific search approaches to find and score matches in selected document search sets. 

We conducted collaborative research with the Multi-Lingual Computing Group at the Army Research Laboratories (ARL) under the Advanced Decision Architectures (ADA) Collaborative Technology Alliance (CTA).  In that research, we prototyped and pilot-tested a task-based approach to machine translation evaluation to support Army machine translation in systems like FALCon.