An AI Copilot Quadrupled the Performance of This Wearable Brain-Reading Device
1 min read
Summary
Machine learning has enabled a wearable brain-reading device to quadruple its output, meaning it can control a robotic arm with ease.
Researchers at the University of California, Los Angeles used a cap studded with 64 electrodes to capture brain signals from outside the skull before feeding them to a machine-learning algorithm.
Using reinforcement learning, the algorithm could predict what the user was trying to do and improve the device’s response, allowing it to complete tasks much faster.
Such non-invasive devices could be used in the future to help people with paralysis or ALS to regain some independence.
However, further iterations of the technology will need to be careful not to override or misinterpret human intent, caution researchers.
