Facebook-Funded Study Can Decode Human Thoughts To Speech In Real-Time

Recently, researchers of the University of California, San Francisco announced that they had successfully decoded human thoughts into speech. This is a milestone for the study that first came to the public notice in April. If things go as the experts expect, we may soon have consumer devices that react to users’ thoughts without them giving audible commands.

Voice control has become the norm when it comes to interacting with smart devices. However, this method cannot be used in public settings or by someone who is mute or has some sort of speech impediment that prevents them from saying the commands as accurately as a normal person. For these situations, communicate by thoughts is by far the best solution.

A few months ago, the researchers announced a neural speech prosthesis that was able to produce something sounds really close to human-made speech from brain activities. If this technology can make it to reality, people with speech impediment may have a chance to speak.  However, what they introduced in April worked too slowly for these usages.

That initial study has become the foundation for today’s achievement as researchers said that they could now decode thoughts in real-time. The vital feature enabling the system to speed up the process is the context in which the person is talking.

David Moses, Ph.D., the lead researcher of this study explained:

This study has the funding of Sponsored Academic Research Agreement with Facebook. For the medical field, the technology may enable people with ALS and the like to communicate with only their thoughts. For Facebook, it may use this breakthrough on the company’s brain-controlled augmented reality glasses.