Scientists from the AIRI Institute for Artificial Intelligence have taught AI to recognize characters in John Ronald Reuel Tolkien’s books by their lines and descriptions, and to determine the nature of the relationship between them. This was reported to socialbites.ca in AIRI.
For training, the researchers took the work of John Ronald Reuel Tolkien and the notes published after the author’s death and edited by his son. The list includes History of Middle-earth, Lord of the Rings, and The Hobbit.
With the help of markers, the researchers extracted 156,482 sentences from the text and retrieved a list of 518 names, 15 race tags, and biographical facts in “manual tuning” mode. After the list was refined, the total number of items for analysis was 880.
Later, the scientists explained the algorithm for automatic analysis of the artifacts. The set of steps included in the work can be used to solve a number of practical problems. For example, analyzing regulatory documents and summarizing legal texts in order to rewrite complex documents in a language that is simple and understandable for an ordinary person.
The technology is suitable for training speech bots and translators, and will also help you save time when searching for information in large volumes of texts.