This technology will help people who are unable to speak to “communicate” in the future.
Scientists from Moscow State University have learned to recognize the sounds of Russian speech during their “internal” pronunciation (to themselves). The experiment was carried out using an electroencephalogram. Such programs for brain-computer interfaces are already being developed abroad. They are of great help to those suffering from neurodegenerative diseases. Until recently, there was no such technology in Russia.
As the “MK” was told at the Department of Information Security of the Faculty of CMC, Moscow State University, now we also have a program for recognizing Russian phonemes. Thanks to specially developed algorithms, the specialists of the department were able to recognize the peculiarities of the internal pronunciation of sounds according to the electroencephalogram data (a method for diagnosing the state of the nervous system, based on recording the bioelectric potentials of the cerebral cortex). It allows you to register specific activity associated with the process of speaking “to oneself”. The work used the methods of probability theory, mathematical statistics and machine learning theory. Experts managed to classify Russian phonemes by electroencephalogram. And now this “decoder” program is not inferior to Western counterparts that decode “thoughts”, or rather, the phonemes of the English language.
At the moment, according to the developers of the algorithm, the results of the work cannot be used in practice, but the ultimate goal is to create a full-fledged device that could help people with speech impairments. In the event of such a device, people would be able to interact with computers using their thoughts.
Newspaper headline:
Dumb people will be able to talk with their brains
.