Building intelligent systems evaluation

Challenges

 

In the implementation of expert systems in Artificial Intelligence, the biggest challenge is to capture expert knowledge for deployment in an automated system that is able to replicate the behaviour of the expert in a process of complex decision-making.
Traditional systems based on rules cover part of the problem but have a poor performance when interpreting the greyer areas of the problem.


Approach

 

EVA is a product conceived, designed and developed by Grupo AIA, as a result of one of the main research lines of the company: to transform information into knowledge.

In this sense, EVA has been specifically designed as a general tool to aid in process automation for complex decision-making.
EVA is a supervised learning tool that captures how an expert makes decisions through dialogue. The expert defines a hierarchical tree of variables, a relevant concept in the decision process and transmits this knowledge through a binary selection process (one of two situations). As a result, EVA provides a concept linear discriminant with the relative weights of the variables which, when applied in qualifying and evaluation of large data volumes, mimics the expert.


Benefits

 

  • Uses supervised learning techniques to be a product of artificial intelligence.
  • Tackles the problem of assessment learning criteria in a process of complex decision-making.
  • Allows definition of the parameters or variables of the decision problem easily.
  • Evaluates and classifies situations, using the criteria of the decision maker.
  • Essential in the process of knowledge acquisition and construction of expert systems for decision-making.
  • Flexible in terms of updating or changing the criteria through a simple and interactive process of review and revision of previous learning processes.

 


Intellectual Property

 

EVA (Trademark)

Ownership: AIA

Appl. No.: 573.538 & M1.596.959 -2