Attention! BrainGPT can read your mind
Vita gazette – BrainGPT systems have become portable. Scientists have developed the world’s first artificial intelligence that can read minds and convert brain waves into texts. This artificial intelligence, which works with a visor covered in sensors, observes certain electrical activities in the brain of the person wearing the visor. At the same time, ıt thinks and converts them into words.
Scientists at the Sydney University of Technology have developed the world’s first artificial intelligence that can read minds and convert brain waves into text. They presented their results at NeurIPS 2023, which took place this week in the US. Western Sydney University’s ICNS team collaborated with partners across the neuromorphic field in developing this ground-breaking project, with researchers from the University of Sydney, the University of Melbourne, and the University of Aachen, Germany.
High-performance portable multichannel wireless EEG acquisition systems can read the minds of willing volunteers. Moreover, brain signals can be converted not just into text but also into speech and expressions that can be rendered using a digital avatar. And these breakthroughs will help people who would otherwise be unable to communicate.
This artificial intelligence, which works with a visor covered in sensors, observes certain electrical activities in the brain of the person wearing the visor. At the same time, ıt thinks and converts them into words. The team that developed this revolutionary artificial intelligence says the invention represents an important step towards treating patients who cannot speak due to stroke or paralysis.
Thought-to-text conversion relies on associating patterns of brain activity with particular words and phrases. And there are a few ways of gathering that information. Notably, the University of Technology Sydney (UTS) team has performed the thought-to-text conversion using a non-invasive headset with dry electrodes that sit over the wearer’s hair.
The wearable EEG measurement device used in their BrainGPT system transmits its 24-bit high-resolution sampled signals, which can be gathered from up to 64 channels via Wi-Fi and features a Cortex-M4 microcontroller.
One step ahead
Volunteers used to lie inside an FMRI machine and listen to podcasts before large language models (LLMs) could decode their thoughts. But not anymore. Portable, high-performance multichannel wireless EEG acquisition systems are capable of brainGPT-enabled thought-to-text conversion. As a video reveals, BrainGPT can turn thoughts (measured by wearing a cap of non-invasive electrodes) into words (shown on a screen). Historically, brain-computer interfaces with high signal quality have been heavy and oversized, but researchers have shown that it’s possible to build a portable setup. Speech recognition systems need to be trained for hours by each user before they generate sensible results, whereas today, they work well straight out of the box.
A few steps later
The team reports that the translation accuracy score is currently around 40% on BLEU-1, a scale between zero and one used to compare the similarity of machine-generated results to high-quality references.
Modern speech recognition systems have BLEU-1 scores closer to 90%. And while it’s clear that the brain’s GPT system has some way to go, it’s worth recalling. Considering the progress made by a raft of neurotech start-ups, the prospects for thought-to-text conversion are very bright. Already, companies are using relatively crude brain wave measurements to inform users about their performance levels and alertness.
Sono stati condotti esperimenti con 29 partecipanti per testare la nuova tecnologia. Il team ha affermato che il tasso di successo dell’intelligenza artificiale è attualmente intorno al 40%, ma continua a lavorare per aumentarlo al 90%.
Professor van Schaik explained that this will lead to advances in smart devices, such as mobile phones, sensors for manufacturing and agriculture, and less power-hungry and more intelligent AI applications. It will also better understand how a healthy or diseased human brain works.
The supercomputer is aptly named DeepSouth, paying homage to IBM’s TrueNorth system, which initiated efforts to build machines simulating large networks of spiking neurons, and Deep Blue, the first computer to become a world chess champion. The name is also a nod to its geographical location.
DeepSouth will be based at Western Sydney University and is crucial to the region’s growth as a high-tech hub.