Sure, looking for news stories with a search engine can tell you about the past, but a new application might act more like a very comprehensive crystal ball. A search engine prototype created by Yahoo's Barcelona lab lets users look into the future -- sort of. The application was revealed this week as part of the Human-Computer Interaction and Information Retrieval challenge , which uses more than 1.8 million New York Times articles from 1987 to 2007 as its data set. The application is also part of the European project LivingKnowledge .
The Yahoo researchers call their creation "Time Explorer" and describe it as "designed for analyzing how new changes over time." Time Explorer expands on our current use and understanding of search engines.
"First, Time Explorer is designed to help users discover how entities such as people and locations associated with a query change over time. Second, by searching on time expressions extracted automatically from text, the application allows the user to explore not only how topics evolved in the past, but also how they will continue to evolve in the future," the researchers wrote in their HCIR challenge paper). "Finally, Time Explorer is designed around an intuitive interface that allows users to interact with time and entities in a powerful way."
Time Explorer goes beyond our current notions of news search engines, which for the most part retrieve stories in reverse chronological order, giving relevance to those stories that are most recent. Currently, users are unable to ascertain a broader understanding of a news story.
So Time Explorer seeks to remedy that by placing news stories on a timeline that reaches both into the past and the future.
"When searching about a regional conflict, for example, a user should be able to identify what factors lead to the conflict, which people where most influential and when, and how the conflict is likely to evolve in the future," the researchers wrote.
Time Explorer extracts temporal information that includes metadata, like when the story was published, and expressions from within the text, like for instance statements about future events. The application also identifies other entities like names of important figures.
"The application includes many features that, in combination, we believe improve upon what is currently available in news search. Most notably, the tight integration between the trend graph, the topic timeline, and the entity list and the ability to search into the future, but also a user interface which allows for easy query renement while still providing visual clues that allow the user to understand how he arrived at the current state," the researchers concluded.
