Between 12 and 14 November, the Institute of Linguistic Research of the Russian Academy of Sciences held the international conference “Linguistic Forum 2020: Language and Artificial Intelligence”. Among the participants of the conference were many Russian and foreign scientists. They discussed contemporary problems of research conducted at the intersection of linguistics and artificial intelligence, including natural language processing.
The researchers were focused on such important topics as language production and understanding, methods of machine learning, linguistic technologies of the future, and the application of neural networks to solve artificial intelligence tasks.
Olga Donina, Associate Professor at the Department of Theoretical and Applied Linguistics at the Faculty of Romance and Germanic Philology of Voronezh State University, gave a talk at the session on “Neural networks and machine learning”. The title of the talk was “Possibility to use machine learning methods to autodetect dead metaphors”. Her speech triggered a heated discussion among the audience. The detection of dead metaphors is an important step in detecting irony and humour in the automatic processing of a natural language as it makes it possible to extract meaning based on language collisions and to eliminate contextual ambiguity, which is absolutely crucial for text based interpretation.