ASERYLA: intelligent agent to emulate the semantic memory through the natural language

The aim of this project is to describe a possible model of how the human memory represents the information relevant to the objects of the world.

Although the brain integrates a lot of information from several sources and types, this model is only focused in the data received in language format.



This is an open software project, and in the menu items you will find:

the neurological basis which this model is based on
how can be emulated in a computer
a detailed model of how is implemented
getting a functional application and their source code
or analyzing the meaning of sentences



This project is be able to process natural language sentences (NLP) to populate a model that stores every general concept plus their characteristics and associations (semantic memory);

though it's not cover the relation between concepts (as it's outside of the scope), this system will provide a core for implement higher mental functions as logic, deduction, language translation,

or even helping to achieve better semantic and pragmatic language level analyzers.

The semantic analysis of sentences can't be processed correctly without specific information about every object mentioned in the phrase,

and that is exactly the functionality that provides the semantic memory.


Take as example the sentence the cat shot the mouse;

   - despite every word exist in the English dictionary, the syntax analysis is correct;

   - the structure fits perfectly with the grammatical English rules [determiner noun verb determiner noun];

   - but this sentence, in a real context (is valid in dreams or an imaginative context), it does not have meaning.

     Because for shoot is needed a gun, for use a gun are necessary a hand, and cats does not have hands.


Contact: ErnestoLB@protonmail.com       about me


Website created on December 17, 2013 (Latest update on November 7, 2019)

www.000webhost.com