Our goal is to make useful linguistic descriptions from data which is obtained by means of sensors. This way, we are trying to define an efficient relationship between a computer's physical world perception (which is potentially able to manage huge data transactions in milliseconds) and its possible uses for a person with well-defined goals and abilities.

Figure outlines the main structure of our proposal. Essentially, the idea is to build a stack of descriptions of the data obtained by sensors, each of them corresponding to different abstraction layers.

Next figure shows a prototype, a device that must be fixed in the car and that, using acceleration sensors, it is able to provide (displaying words on the LCD) a linguistic description of the car driver's style of handling.

Top

The meaning of a Natural Language clause is determined by:

• Syntactic and semantic aspects extracted considering Natural Language as a way of coding the information.
• Spatial and temporal context where the message transaction is happening. Including the role of each partner.
• The beliefs, desires and intentions of communication partners.
• Emotions and feelings which are released during the communication process.

This line of work focuses on finding the most suitable data structures and algorithms for the processing of the meaning of words during the communication between a human and a computer using natural language.

As a practical application of our efforts in this line, we have joined forces with a Savings Bank and an industrial company in the project of creating an improved version of Automatic Teller Machines.

Top

One of the main foundations of Computation with Meanings is Zadeh's proposal about “Computation with Words and Perceptions”.

The goal in this line of work is to study the possible extensions of Fuzzy Logic and related tools to the open and complex domain of the meanings of the words used when we describe the real world using natural language.

Top