Paper Review 1: ELIZA — A Computer Program for the Study of Natural Language Communication Between Man and Machine
In this post, the paper “ELIZA — a computer program for the study of natural language communication between man and machine” is summarized.
Joseph Weizenbaum. 1966. ELIZA — a computer program for the study of natural language communication between man and machine. Commun. ACM 9, 1 (January 1966), 36–45
Eliza is a computer program performing primitive forms of natural language processing tasks. It was created by Joseph Weizenbaum at MIT in 1966. It makes natural language conversations between a human and a computer possible. Eliza enables communication with pattern matching and substitution approaches but it does not use context information. Still, it is an important step in the evolution of natural language processing. Eliza processes user inputs and produces output for a given input, based on the rules provided by scripts. The creator of Eliza, Weizenbaum, identified 5 problems to be solved in Eliza scripts. These are the identification of key words, the discovery of minimal context, the choice of appropriate transformations, generation of responses in the absence of key words and the provision of an editing capability for Eliza scripts. The answers to these questions are critical in forming scripts and Weizenbaum solved these problems.
Eliza’s input processing starts with key words analysis. Key words are important and scripts assign ranks to these key words. If there are key words, they are sorted according to their rank and the input sentence is processed using the rule associated with the key word having highest rank. After identifying key words, Eliza applies a transformation rule, consisting of two parts — decomposition and reassembly. Decomposition rule breaks the input into pieces according to key words and applies proper transformations. Then, decomposition rule forms reassembly rules to follow when reconstructing the sentence. This rule takes parts of the input sentence that the decomposition rule created and rearranges them in a suitable form. All in all, Eliza implements a bunch of rules to create response from a user input.
