0

2 billion images/sec: brain-like chip developed

Share

Skynet is getting closer and closer

Researchers at the University of Pennsylvania have developed a powerful new optical chip that can process nearly 2 billion images per second. The device consists of a neural network that processes information in the form of light without the need for “components that slow down traditional computer chips, such as memory.”

At the heart of the new chip is a neural network, a system modeled after how the brain processes information. These networks are made up of nodes that interconnect like neurons, and they even “learn” like an organic brain by being trained on datasets. Over time, they become much better at their tasks.

Check:  Students Develop Innovative Wheelchair Case: Here's How It Works

But instead of electrical signals, the new chip processes information in the form of light. As neurons, it uses optical wires stacked in several layers, each of which specializes in a certain type of classification.

In testing, the team fabricated a 9.3mm2 chip and had it classify a series of handwritten letter-like characters. After training on the appropriate datasets, the chip was able to classify images with 93.8 percent accuracy for sets containing two types of characters and 89.8 percent for four types.

Most impressively, the chip was able to classify each character within 0.57 nanoseconds, allowing it to process 1.75 billion images per second. The team says the speed is due to the chip’s ability to process information in the form of light, which gives it a number of advantages over existing computer chips.

Check:  Ferrari Purosangue unveiled: the new $390,000 crossover king

Another advantage is that the information being processed does not need to be stored, so the chip also saves time by not having to send data to memory, as well as physical space, since the memory does not need a component at all. The team also claims that not storing data is also more secure, as it prevents potential leaks.

To Read Great Articles, Click Here
Follow Us On Facebook Twitter Telegram