News
Facebook Is Making A Machine That Will Decode Speech Directly From Your Brain
Finally, the first step.
Facebook declared that they were developing a non-invasive technology that will achieve speech decoding at 100 words per minute directly from the brain. Now, we’ve got a great update on the research.
TL;DR, no, they haven’t achieved that, but the first step of decoding has been fulfilled.
“The promise of AR lies in its ability to seamlessly connect people to the world that surrounds them – and to each other. It’s a tantalizing vision, but one that will require an enterprising spirit, hefty amounts of determination, and an open mind,” says ambitious researchers on a Facebook blog.
Don’t forget a lot of capital money to invest in the research (which Facebook definitely has no problem with).
The advancement of the technology, however, relies on other fields in relation to understand the complexity of the brain and integrating it into technology. Particularly, BCI or brain-computer interface research which still require invasive techniques such as implanting electrodes in our brain.
In a Facebook-funded study by University of California, San Francisco, epilepsy patients with normal speech and otherwise healthy conditions participate. They were to answer simple questions with fixed choices as answers. For the first time in the history of BCI, researchers were able to “decode a small set of full, spoken words and phrases from brain activity in real-time – a first in the field of BCI research.”
Facebook says that it’s still a long way to achieve this and their end goal of 100 words per minute with less than 17% of error with non-invasive techniques. Another way researchers have been trying is using near-infrared light.
“Like other cells in your body, neurons consume oxygen when they’re active. So if we can detect shifts in oxygen levels within the brain, we can indirectly measure brain activity,” explains Facebook. Yet, they also claimed that this won’t lead them to specific answers and only recognize simple words like ‘home’, ‘select’, and ‘delete’.
