Foundation concept: Artificial intelligence.

Definition: The development and use of techniques to understand natural language input and to formulate an intelligent response.

Overview

Since the earliest days of computing, the idea of interacting with a machine through speech has been very appealing. To do so requires solving two problems: firstly, a system must be able to take a captured audio signal and recognize the words within it; and secondly, the system then has to be able to understand what the whole thing means. Researchers have been working upon both of these problems and have achieved a certain degree of success in each. It is not uncommon for business users to interact with airline reservation systems, for example, through an automated voice-recognition system. Many of the more successful systems bypass most of the first stage by accepting only typed input, not audible speech.

The origins of Natural language processing (NLP) are almost as old as computing itself. In the 1950s and 1960s researchers began to develop a variety of techniques aimed at understanding the ways computers could be used to provide NLP. One such system, Eliza, was created by Joseph Weizenbaum, whose system pretended to be a psychologist with whom a dialog could be enacted; a typical interaction with Eliza might start like this:

1. Eliza: How are you?

2. Human: I'm feeling ill.

3. Eliza: How long have you been feeling ill?

4. Human: Just a few days.

5. Eliza: Do carry on.