Introduction
Although no one definition is universally accepted, Artificial Intelligence is generally defined as the field interested in creating automated systems that behave in intelligent ways. By intelligent behavior we mean behavior that simulates that of humans. The hope is that more intelligent computers will be more helpful to humans (by understanding our needs better, and being able to satisfy them more efficiently), while at the same time helping us understand how it is that people perform intelligent tasks. One of the most complicated tasks performed by humans is that of Language Processing. Because of this, an effort has been made to develop computer systems able to process natural languages (such as Spanish, English, or German). A computer system of this type would be able to perform tasks such as understanding spoken instructions, understanding instructions written in a natural language, or replying to queries by producing natural language sentences. A system that is able to process natural language could also provide information about how it is that humans learn and process language. One approach used in trying to develop this type of system is that of Artificial Neural Networks (ANN). Different researchers have used ANNs to solve language problems, with varying levels of success. The fact that ANNs can be configured in any of several different ways raises the following question: Are any Neural Networks able to solve natural language problems, or are there key configuration parameters that have to be set in a particular way in order to have a successful NN? In the following pages I will provide a quick introduction to the field of Artificial Neural Networks and Natural Language Processing and outline how I intend to explore the above question.

Introduction

Although no one definition is universally accepted, Artificial Intelligence is generally defined as the field interested in creating automated systems that behave in intelligent ways. By intelligent behavior we mean behavior that simulates that of humans. The hope is that more intelligent computers will be more helpful to humans (by understanding our needs better, and being able to satisfy them more efficiently), while at the same time helping us understand how it is that people perform intelligent tasks.

One of the most complicated tasks performed by humans is that of Language Processing. Because of this, an effort has been made to develop computer systems able to process natural languages (such as Spanish, English, or German). A computer system of this type would be able to perform tasks such as understanding spoken instructions, understanding instructions written in a natural language, or replying to queries by producing natural language sentences. A system that is able to process natural language could also provide information about how it is that humans learn and process language.

One approach used in trying to develop this type of system is that of Artificial Neural Networks (ANN). Different researchers have used ANNs to solve language problems, with varying levels of success. The fact that ANNs can be configured in any of several different ways raises the following question: Are any Neural Networks able to solve natural language problems, or are there key configuration parameters that have to be set in a particular way in order to have a successful NN? In the following pages I will provide a quick introduction to the field of Artificial Neural Networks and Natural Language Processing and outline how I intend to explore the above question.

Back to the Table of Conntent

Back to the previous topic

To the Next topic