Mind reading? AI interprets thoughts through brain waves

A science and technology advance by leaps and bounds, bringing to light extraordinary achievements. This time, Australian researchers ushered in an era in mind-machine interface with DeWave.

It is an artificial intelligence (AI) capable of deciphering silent thoughts directly from brain waves.

Thus, DeWave is a pioneering AI system, developed by scientists at the University of Technology Sydney (UTS), that has the power to transform brain waves into text.

Surprisingly, this feat is achieved without the need for invasive surgery, just with the use of a “comfortable cap” that records brain waves using an electroencephalogram (EEG).

Testing and the search for accuracy

During tests carried out with more than two dozen participants, DeWave achieved an initial accuracy of just over 40%.

Despite this, this achievement represents a significant advance, increasing accuracy by 3% compared to previous methods of translating thoughts from EEG recordings.

The researchers’ ambitious goal is to achieve an accuracy of around 90%, comparable to conventional language translation methods or speech recognition software. See the video about the tests below:

Unlike other methods that require invasive implants or expensive technologies, DeWave is non-invasive and offers a new horizon in communication, especially for patients with stroke, paralysis and even for the operation of machines such as bionic arms or robots.

However, challenges persist, especially in the translation of nouns, as the semantic nuances of these terms prove complex for the system.

The future of neuroscience and AI

The scientists behind DeWave are optimistic about the future of this technology and foresee continued advances.

Integration with large language models and refinement of the coding process promise to open new frontiers in neuroscience and AI, which will revolutionize our understanding of the interaction between the brain and language.

The research, although showing significant progress, requires further improvement, especially regarding the issue of noise in the EEG signal captured through the cap, compared to electrodes implanted in the brain.

The future looks promising, with the potential to make direct translation of human thoughts into the day to day.