Mind-reading AI can convert brain waves into text

IA que lee mentes

A system that monitors the brain’s electrical activity through the scalp has the ability to transform thoughts into words by implementing a large language model, although the results are still far from being completely accurate.

Through the exclusive use of a helmet equipped with sensors and powered by artificial intelligence, a group of scientists has reported their ability to translate a person’s thoughts into written text .

In the study, participants read text excerpts while wearing a cap that recorded the electrical activity of their brains through their scalp. Subsequently, these electroencephalogram (EEG) recordings were converted into text using an artificial intelligence model called DeWave .

robot multiproposito
robot mas pequeno del mundo
agua en la superficie de 2 asteroides
robot enviado a la EEI
robot submarino 1
ADAM: Meet the multipurpose professional assistance robot for homes
Students built the world’s smallest robot
Scientists find water on the surface of two asteroids
Robot sent to the ISS can perform operations in space
Autonomous underwater robot developed to combat biofouling on ships
previous arrowprevious arrow
next arrownext arrow

DeWave: the AI ​​that reads minds

Chin-Teng Lin, from the University of Technology Sydney (UTS) in Australia, highlights that the technology in question is non-invasive, has a relatively low cost and is easily transportable. Although the system is far from perfect, with an accuracy hovering around 40%, Lin mentions that more recent data, currently undergoing peer review, indicates an improvement in accuracy, exceeding 60%.

In the study presented at the NeurIPS conference in New Orleans, Louisiana, participants read aloud, although the DeWave program does not rely on spoken words. However, in the team’s latest research, participants read the sentences silently. In the previous year, a team led by Jerry Tang of the University of Texas at Austin reported similar accuracy in converting thoughts to text, but they used MRI scans to interpret brain activity. The use of EEG is more practical, since subjects do not need to remain motionless inside a scanner.

Team member Charles Zhou of the University of Technology Sydney (UTS) explains that the DeWave model was trained by observing numerous examples in which brain signals correspond to specific sentences.

For example, when you think about saying ‘hello,’ your brain sends certain signals ,” Zhou says. ” DeWave learns how these signals relate to the word “hello” by seeing many examples of these for different words or sentences .”

After DeWave effectively understood the brain signals, the team integrated them with an open-source large language model (LLM), similar to the artificial intelligence that powers ChatGPT .

This LLM is like a smart writer who can form sentences. We tell this writer to pay attention to DeWave signals and use them as a guide to create sentences”says Zhou.

In the final phase, the team conducted joint training of DeWave and the language model to further fine-tune sentence generation based on EEG data. With further refinement, the researchers anticipate that this system could transform the communication of individuals who have lost the ability to speak, such as those affected by a stroke, and could also find applications in the field of robotics.

Craig Jin of the University of Sydney praises the work of Lin’s team, describing it as outstanding progress.

At the forefront of innovation, artificial intelligence has reached a surprising milestone: the ability to “read minds.” This breakthrough promises to transform our understanding of the human mind and open the doors to previously unimaginable possibilities. Are we ready to explore the limits of the mind with this revolutionary technology?

Don’t miss any of our posts and follow us on social media!

Inspenet.com YouTube LinkedIn Facebook Instagram

Source: newscientist.com

Share this news on your social networks

Rate this post
1 star2 stars3 stars4 stars5 stars (No rating yet)