When we talk about Artificial Intelligence it seems that we always talk about how we can improve the future and make the present more productive, but sometimes it can help a lot to discover more about our past.
It is already used in some branches of archeology, and today I will talk about a system used to discover famous people in ancient documents.
The basic idea is this: thousands of old documents are scanned and used character recognition systems To extract names of people, the result is then analyzed and the number of times each person appears is verified, thus being a symbol that said person was very relevant at the time.
That is the quick and superficial explanation of a new artificial intelligence system created by a researcher at the School of Management at the University at Buffalo, in the United States.
The algorithm has been published in Decision Support Systems, where they comment that it can find and rank people’s names in order of importance from the results produced by optical character recognition. They have partnered with the New York Public Library and analyzed more than 14,000 articles from the city’s Sun newspaper, articles published during November and December 1894.
Once scanned, people’s names are ranked by importance based on a number of attributes, including the context of the name, the title before the name, the length of the article, and how often the name is mentioned in an article .
The algorithm learns these attributes from text only, without relying on external sources, and then statistical measures are used to classify the people found.
In this way you can learn more about important people during a specific one, and in the future it would even be possible to examine the relationships between people to better understand the story from the automatic analysis of what was written.