Artificial intelligence, a helmet that reads minds


Right in thought and then converts words think in text. This is the purpose of a new experimental device from Artificial Intelligence and just developed by researchers at the University of Technology Sydney. But the novelty of the system, which its creators presented at the annual congress NeuroIpsis that it is portable and non-invasive and for this reason, it could offer new hope to people who can no longer because of illness or injury communicate neither verbally nor with gestures or eye movements.

Brain-computer interface

In recent years, more and more technology proposals have appeared and brain-computer interface who can convert brainwaves into lyrics. However, most of these systems are invasive and require surgery to implant them electrodes in the user’s brain or are only capable of basic yes/no communication. For example, more than a month ago, researchers from Duke University reported on the site The nature of communication that he developed a device equipped with 256 microscopic brain sensors on a postage stamp-sized fragment of material. And while their prototype proved faster than those being tested, it still required electrodes to be implanted into patients’ brains.

Non-invasive system

Otherwise, the new device requires users to wear a Eeg helmet (electroencephalogram) connected to a computer. The DeWave software used by the device was trained by recording and analyzing the electrical signals produced by the device brain a total of 29 volunteers while silently reading some passages of text. In other words, DeWave’s AI-based algorithms learned which specific EEG signals corresponded to specific written words and phrases. It was then able to detect these signals when the user was not reading any words, but only thinking about them.

Next step

Currently, its level of accuracy is approx 40% according to the Bleu scale (BiLingual Evaluation Understudy), which is precisely the measure of accuracy for automatically translated text. Now the team is already working on further improving this technology and reaching the level accuracy higher, about 90%. “This research represents a pioneering effort in translating raw EEG waves directly into language, marking a major breakthrough in the field,” said study author CT Lin. “He is the first to incorporate discrete coding techniques into the brain-to-text translation process, introducing an innovative approach to neural decoding.”


Source link

Leave a Comment