First Non-Invasive AI System DeWave Can Read Mind, Turn Silent Thoughts Into Texts
There is a way to translate your thoughts into texts without invasive surgeries. Australian researchers developed a new technology using artificial intelligence (AI).
New AI System That Can Translate Silent Thoughts Into Texts
Australian researchers behind the AI system DeWave used data from over two dozen people to evaluate their methodology. With just a tight-fitting cap to wear, DeWave is the first non-invasive artificial intelligence device in the world that can translate silent thoughts into text.
Wearing a cap that recorded their electroencephalogram (EEG) brain waves and translated them into text, participants read silently. Though this is merely a 3 percent gain over the previous standard for thought translation from EEG recordings, DeWave only reached slightly over 40 percent accuracy based on one of two sets of metrics in trials conducted by computer scientist Chin-Teng Lin from the University of Technology Sydney (UTS) and colleagues.
With more development, DeWave may facilitate communication between stroke and paralysis victims and persons operating robotics and prosthetic arms. The researchers want to increase accuracy to about 90%, comparable to speech recognition software or traditional language translation techniques.
After considerable training, DeWave's encoder converts EEG waves into a code, which may then be matched to certain phrases based on how similar they are to entries in DeWave's "codebook."
"This research represents a pioneering effort in translating raw EEG waves directly into language, marking a significant breakthrough in the field," Lin said.
"It is the first to incorporate discrete encoding techniques in the brain-to-text translation process, introducing an innovative approach to neural decoding."
Other Alternatives
Other approaches to converting brain signals into language are unrealistic for everyday usage because they involve intrusive procedures to implant electrodes or large, costly MRI equipment. They also frequently need eye-tracking to translate brain signals into word-level chunks.
It's reasonable to assume that when someone's eyes jump around from word to word, their brains take a moment to absorb each one. Translating raw EEG waves into words without eye tracking to identify the associated word target is more complicated. It is difficult to educate AI to understand individual thoughts since various people's brain waves don't reflect word breaks in quite the same manner.
One of the most talked about technologies in this area is Elon Musk's Neuralink. FDA approved it for human clinical trial earlier this year.
The trial results are far more important than disclosing a potential issue with future humans. But this isn't just any trial. This represents the first instance of a technological connection to the human brain.
In related news, a brain chip technology helped a woman with amyotrophic lateral sclerosis (ALS) to speak again 13 years after she was diagnosed with the condition. Pat Bennett was implanted with four tiny sensors in her brain, which transmit her ideas --from her brain to a computer monitor -- at 62 words per minute, which was faster than the previous speed record for the technology.
Neuralink co-founder and brain-machine interfaces expert Professor Philip Sabes was impressed with the result and deemed it a "big breakthrough."
RELATED ARTICLE: Elon Musk is Looking for a Neuralink Clinical Trial Director; Does It Mean It Has A Toxic Work Environment?
Check out more news and information on Neuralink in Science Times.
