The world's first `mind-reading helmet`... Is it possible?

2023-12-16 2023-12-16T18:10:16Z

Weather of Arabia - Scientists at the University of Technology in Sydney have developed the first artificial intelligence system in the world capable of reading thoughts, as it translates brain waves into texts that can be read. This system, which is considered a revolution in the field of medical care, is done through the use of a helmet that contains a sensor that monitors activity. Electrophoresis is detected in the brain during thinking, and converts it into words.

The university's research team explained that this technology could revolutionize the care of patients who have lost the ability to speak as a result of stroke or paralysis. An explanatory video clip shows a person thinking about a sentence that appears on the screen, and then it is converted into text by an artificial intelligence model. It showed The results are an almost exact match.

The team also believes that this innovation will enable seamless control of devices such as bionic limbs and robots, allowing humans to give directions just by thinking about them.

Achieving a revolution in the science of artificial intelligence

Professor CT Lin, lead researcher, explained that this research represents a major advance in translating EEG waves directly into language, and presents an innovative approach to neural decoding, noting that integration with large-scale language models opens new horizons.

In the fields of neuroscience and artificial intelligence, and in comparison with previous technology for translating brain signals, which required surgical operations to implant electrodes or use of an MRI machine, this new technology relies on a simple helmet that reads the individual’s thoughts.

To test efficiency, the researchers conducted experiments with 29 participants , where they were shown sentences and asked to think about reading them. The AI model then analyzed the brainwaves and translated them into text, and although the translation accuracy is currently around 40%, the team is working hard to improve this to 90%.

This new technology, based on DeWave, demonstrates its ability to translate EEG signals into words using large language models, using the BART model and combining bidirectional BERT context and a left-to-right ChatGPT decoder.

In conclusion, the team is working to improve the accuracy of this system to achieve greater progress in the field of translating ideas into text.