Shulou(Shulou.com)11/24 Report--

CTOnews.com, May 4-A team at the University of Texas at Austin has trained a neural network to decode functional magnetic resonance imaging (fMRI) signals from the human brain by non-invasive decoding.

"We have introduced a non-invasive decoder that uses cortical semantic representations recorded by functional magnetic resonance imaging (fMRI) to reconstruct continuous language," the team wrote at the beginning of the paper, published in the journal Nature Neuroscience (Nature Neuroscience).

The researchers used the GPT-1 model to analyze the data recorded by fMRI to explain the corresponding activity of the human brain and achieve the effect of "mind reading" in the novel.

The biggest breakthrough in this study is to complete the process in a non-invasive manner, while other similar systems currently require surgical implantation. The team also pointed out that participants were not required to use only the words in the prescribed list.

At present, the accuracy of the study is not very high. For example, the brain means "I haven't got my driver's license yet" and GPT-1 translates it as "she can't drive yet."

CTOnews.com attaches a link to the paper here, which interested users can click to read.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.