The new native algorithm can automatically analyze artwork and identify the types of characters and their relationships. This was announced by the AIRI Institute for Artificial Intelligence.
The work, led by Ilya Makarov, was carried out by adapting pre-existing neural network algorithms for text analysis to the chosen task. NER algorithms are used to analyze natural language to separate characters from text and the nature of their relationships.
“The set of steps involved in the work can be used to extract named entities and their relationships from other texts. With the help of the described methods, a number of practical tasks can be performed. For example, analyzing regulatory documents and legal texts to rewrite complex documents in a language that is simple and understandable for an ordinary person. to summarize,” he said.
The operation of the system was tested by John Tolkien on the Lord of the Rings saga. It turns out that the algorithm allows you to recognize 880 entities and understand how they are related to each other. In addition, the neural network was able to recognize them by characteristic cues. The authors of the development believe that with its help it will be possible to automatically prepare a summary of important documents.
previous experiment showedThe disappearing balloon trick can’t fool the monkeys.