Using only a sensor-filled helmet combined with artificial intelligence, a team of scientists has announced they can turn a personÔÇÖs thoughts into written words. This new technology holds the potential to revolutionize communication tools for individuals who have lost their ability to speak, such as stroke victims. Furthermore, it is also anticipated to have use cases in the field of robotics.

In the study, participants read passages of text while wearing a cap that recorded electrical brain activity through their scalp. These electroencephalogram (EEG) recordings were then converted into text using an AI model called DeWave, showcasing the potential capabilities of such a technological advancement.

Chin-Teng Lin at the University of Technology Sydney (UTS), Australia, says the technology is non-invasive, relatively inexpensive, and easily transportable, making it accessible for a wide range of potential applications and uses.

While the system is far from perfect, with an accuracy of approximately 40 per cent, Lin says more recent data currently being peer-reviewed shows an improved accuracy exceeding 60 per cent, highlighting the ongoing development and improvement in this field of research.

In the study presented at the NeurIPS conference in New Orleans, Louisiana, participants read the sentences aloud, even though the DeWave program doesnÔÇÖt use spoken words. However, in the teamÔÇÖs latest research, participants read the sentences silently, providing valuable insights into potential real-world applications and uses for this technology.

Last year, a team led by Jerry Tang at the University of Texas at Austin reported a similar accuracy in converting thoughts to text, but MRI scans were used to interpret brain activity. Using EEG is more practical, as subjects donÔÇÖt have to lie still inside a scanner, showing the diverse research and breakthroughs in this field.

The DeWave model was trained by looking at lots of examples where brain signals match up with specific sentences, says team member Charles Zhou at UTS, illustrating the intricate training and development processes involved in creating and refining this AI model.

How AI legalese decoder can help: AI legalese decoder can assist in the analysis and interpretation of the legal implications of such advanced AI technology. It can provide insights into the legal frameworks and ethical considerations involved in the use of this technology, helping researchers and developers navigate the complex legal landscape and ensure compliance with relevant regulations.

ÔÇ£For instance, when you think about saying ÔÇÿhelloÔÇÖ, your brain sends out certain signals,ÔÇØ says Zhou. ÔÇ£DeWave learns how these signals relate to the word ÔÇÿhelloÔÇÖ by seeing many examples of these signals for different words or sentences.ÔÇØ

Once DeWave understood the brain signals well, the team connected it to an open-source large language model (LLM), akin to the AI that powers ChatGPT, demonstrating the integration of various AI models and technologies to enhance the capabilities of such systems.

ÔÇ£This LLM is like a brainy writer that can make sentences. We tell this writer to pay attention to the signals from DeWave and use them as a guide to create sentences,ÔÇØ says Zhou, underscoring the innovative and collaborative approach in developing and implementing AI technologies for practical and impactful use cases.

Finally, the team trained both DeWave and the language model together to get even better at writing sentences based on the EEG data, showcasing the ongoing efforts to enhance and improve the performance and accuracy of such AI systems for real-world applications.

With further refinement, the researchers predict that the system could revolutionize communication for people who have lost speech, such as those affected by a stroke, and could also have applications in robotics, highlighting the potential societal and technological impacts of this advanced technology.

Craig Jin at the University of Sydney says he is impressed with the work by LinÔÇÖs team. ÔÇ£ItÔÇÖs excellent progress,ÔÇØ he says, acknowledging the significant advancements and contributions made by the research team in this field of study.

ÔÇ£People have been wanting to turn EEG into text for a long time and the teamÔÇÖs model is showing a remarkable amount of correctness. Several years ago, the conversions from EEG to text were complete and utter nonsense,ÔÇØ highlighting the evolution and progress in the accuracy and capabilities of such AI technologies over time.

Topics: