NewsFinder: Automating an AI News Service

week about AI, or any other topic, requires more than searching for individual terms. Since it is time-consuming to find and post interesting stories manually, we have designed and written an AI program called NewsFinder1 that automatically collects English-language news stories from the web, categorizes them with trained models, filters out uninteresting and unrelated stories based on a set of heuristics, and publishes them in summarized form as the AI in the News service. AI in the News is part of the AAAI AITopics2 site described previously (Buchanan, Smith, and Glick 2008). The goal for NewsFinder is to publish a small, select set of news stories that are of general interest to the AI community and to provide categorization metadata to help readers scan for stories that match their specific interests. A news story may be assigned to a single category or to multiple categories (for example, Robots and Vision). The task is similar to asking Google News3 to find stories about AI, but differs in one important respect. Google News is driven by readers’ queries to find stories containing a set of keywords from thousands of news sources. By contrast, NewsFinder uses multiple predetermined queries and Really Simple Syndication (RSS) feeds to find stories from a preferred set of hand-selected sources, as well as from Google News query searches. Google News can be described as a “pull” service because readers must perform queries in order to find news stories that match their interests. NewsFinder, on the other hand, is a “push” service that selects news stories based on Articles