Mind reader: This revolutionary AI system can translate silent thoughts into text

Researchers from the GrapheneX-UTS Human-centric Artificial Intelligence Centre at the University of Technology Sydney (UTS) have come up with a unique concept as they have developed an AI-equipped cap (or helmet) that can apparently read thoughts. The revolutionary AI system is capable of translating silent thoughts into text by using electroencephalography (EEG).

The system, termed as DeWave, isn't perfect, but its capability of roughly translating internal thought processes into textis remarkable.

The researchers stated that the system has an accuracy of around 40 per cent. Meanwhile, more recent data, being peer-reviewed, reveals an enhanced accuracy of more than 60 per cent. Hence, further improvement is expected.

Add WION as a Preferred Source

The research was led by professor Chin-Teng Lin, first author Yiqun Duan, and PhD candidate Jinzhou Zhou, and the findings can be significant for those who can't speak due to stroke or paralysis.

DeWave consists of two major components - an EEG-reading cap and a large language model (LLM).

Participants of the study read passages of text while wearing a cap that recorded electrical brain activity through their scalp. These EEG recordings were then converted into text using the AI model.

As quoted by NewsScitist, Chin-Teng Lin at the University of Technology Sydney (UTS), said that the technology is non-invasive. He claimed that it is relatively inexpensive and easily transportable.

The study was presented at the NeurIPS conference in New Orleans, Louisiana, where participants read the sentences aloud, despite the fact that the DeWave system doesn't use spoken words. However, the participants read the sentences silently in the team's latest research.

According to team member Charles Zhou at UTS, the DeWave system was trained by looking at a large number of examples. They studied brain signals and how they match up with certain texts.

As quoted, Zhou said, "For instance, when you think about saying 'hello', your brain sends out certain signals." He said that the AI system learns how these signals "relate to the word 'hello' by seeing many examples of these signals for different words or sentences".

(With inputs from agencies)