Scientists, leveraging artificial intelligence and large language models, have achieved a breakthrough in mind-reading capabilities.
As reported by NewScientist, a team of researchers has unveiled a technology that can translate a person’s thoughts into written words using only a helmet equipped with sensors and artificial intelligence.
In this study, participants wore a hood while reading text passages, and the electrical brain activity recorded by the electroencephalogram (EEG) sensors on their scalp was subsequently converted into text using an artificial intelligence model named DeWave.
The DeWave model’s brain signals were trained by analyzing a substantial number of examples, matching specific sentences to corresponding signals.
For instance, when a person thinks about saying ‘hello,’ certain signals are sent by the brain. By exposing DeWave to numerous instances of these signals for various words or phrases, the model learns the correlation between these signals and the word ‘hello.
‘ Once DeWave comprehends the brain signals, the team integrated it with an open-source large language model (LLM), akin to the artificial intelligence underlying ChatGPT.
This LLM acts as an intelligent writer capable of constructing sentences, guided by the signals from DeWave to form coherent sentences.
Through joint training of DeWave and the language model, the researchers enhanced their ability to generate sentences based on EEG data.
The researchers anticipate that as the system undergoes further development, it will significantly facilitate communication for individuals who have lost the ability to speak, such as those recovering from a stroke.
This groundbreaking technology holds the potential to provide a valuable communication tool for patients facing speech-related challenges.
You may also like this content
- Meta Building World’s Fastest AI Supercomputer for Metaverse
- Artificial Intelligence Will Make Decisions Instead Of People
- Bill Gates: Artificial Intelligence Over Web3 and Metaverse
-
-
-
-
-
-
-
-
-
-
