Reading the mind is one of the fascinations of the human being because it is perhaps the most unknown or still undeciphered and understood part of the body. Artificial intelligence has set that challenge and has discovered how it works.
The University of California in San Francisco has developed a system capable of reading minds, translating thoughts and making them speak, all without moving your mouth.
The system, built on artificial intelligence and machine learning, repeats translated phrases aloud directly from the brain. To do this, the system has learned to translate a series of impulses from the human brain.
The first experiments have made repeating a sentence of 250 words possible. Through a test the system harnessed electrodes implanted in the brains of these patients to record brain activity as they spoke the sentences.
The project is known as Brain-to-Speech; the goal of the project is to develop a system that can translate the language of the brain into speech in real time without the need for any physical movement. The technology could have significant implications for people who have lost their ability to speak due to neurological disorders or injuries.
How Brain-to-Speech works
The Brain-to-Speech system works by recording the brain’s electrical activity using electrodes implanted in the patient’s brain. A machine learning algorithm then analyzes this electrical activity to identify patterns corresponding to speech. The algorithm then uses these patterns to produce synthesized speech in real-time, which is played back to the patient through a speaker or headphones.
The initial tests of the Brain-to-Speech system were conducted on epilepsy patients with electrodes implanted in their brains to monitor seizure activity. During the tests, the patients were asked to read aloud from a set of predetermined sentences while their brain activity was recorded.
These data were fed into a neural network developed to identify patterns that could be associated with the way humans speak.
click here to read : Next Generation Computing : The Future of Technology
The system is complemented by a second neural network that takes advantage of the patterns discovered by the first to create words from a sentence.
Once the algorithm had been trained, the researchers tested the system by asking the patients to think of sentences that were not included in the original set. The Brain-to-Speech system was able to translate these novel sentences into synthesized speech accurately, demonstrating the system’s ability to adapt to new patterns of brain activity.
However, researchers have to overcome the unpredictability of the human being because it is difficult to repeat two similar phrases. So the system has to learn to interpret different sentences. While the Brain-to-Speech system is still in the early stages of development, and it is one of the best projects in terms of progress and success, there is still a lot of work to correctly interpret what the human mind says.