STEM与日常科技·英语30篇（3）

1 / 30

神经形态芯片与事件驱动计算

Unlike traditional chips that run at fixed clock speeds, neuromorphic chips activate only when new data arrives.
Each tiny 'neuron' on the chip fires independently, just like real brain cells responding to sensory input.
This event-driven design cuts power use by up to 90% compared with standard processors in smart sensors.
Cameras using this tech ignore static scenes but instantly process motion—ideal for security or robotics.
Because computation happens locally and asynchronously, delays from sending data to the cloud disappear.
Researchers now embed these chips into hearing aids that adapt in real time to noisy rooms.
No background clock means less heat, longer battery life, and quieter operation for wearable devices.
While still emerging, this architecture helps machines sense and react more like living organisms do.