An Architecture for Distributed Natural Language Summarization

We present a system that incorporates agentbased technology and natural language generation to address the problem of natural language summarization of live sources of data. The input to the system includes newswire and on-line databases and ontologies. The output consists of short summaries that convey information selected to fit the user's interests, the most recent news updates, and historical information. The system is under development. 1 Introduction One of the major problems with the Internet is the abundance of information and the difficulty for the average computer user to read everything existing on a specific topic. There exist now more than 100 sources of live newswire in operation on the Internet. The user has to go through megabytes of news every day to select articles of interest and read the relevant parts of them. Hence, he needs search and selection services, as well as for summarization facilities. There currently exist more than 10 operational search and select..