Who said what: using machine learning to correctly attribute quotes

Dec 1, 2023 ·

Machine learning algorithms are being used to help verify and attribute quotes more accurately. Quotes attributed to a particular person can often be wrong, especially in cases where the source material is old or difficult to locate. In recent years, machine learning algorithms have been developed to help identify and correctly attribute quotes.

At the University of Manchester, researchers developed a system called QUOTE-ML that uses machine learning techniques to analyze text taken from books, articles, speeches, and other sources. The system can then compare the source material to a database of known quotes and evaluate how closely it matches them. By comparing the results with existing quotes, it can determine if the quote should be attributed to a particular person.

The team tested the system's accuracy on a set of 9,000 previously attributed quotes, and found that it was able to correctly attribute quotes with an accuracy rate of up to 95%. The system also outperformed existing methods of quoting detection, such as manual searches, by at least 10%.

One of the benefits of this system is that it can help distinguish between quotes that were wrongly attributed and those that were actually said by the relevant person. This is particularly useful for quotes that are more than a few decades old, since it can be difficult to track down reliable source material.

Overall, the use of machine learning techniques to accurately identify and attribute quotes is a major development in the field of linguistics. By using these algorithms, it is now possible to verify the accuracy of quotes attributed to famous people, which helps to ensure that their words are properly remembered and understood.