AI Designed Computer Chips That the Human Mind Can’t Understand – Popular Mechanics
- A new neural network process has designed wireless chips that can outperform existing ones.
- This convolutional neural network analyzes the desired chip properties then designs backward.
- Much of AI news is hype, but this is open access, peer reviewed research in a reputable journal.
Our world runs on computer chips. From the chips that run new cars to the chips that help your phones and computers process information to the microchips that help track lost animals, there’s very few aspects of modern human life that are not touched by chips. As a result, there’s a huge and consistent push to make better and more innovative chips as fast as possible—sometimes by any means necessary. And apparently, that sometimes means taking a portion of the design control out of human hands.
A group of scientists recently explained their process of letting artificial intelligence technology (AI) design and test a more efficient computer chip. The lead author—electrical engineer Kaushik Sengupta of Princeton University’s Sengupta Lab—was recently awarded an IEEE fellowship for his wireless chip and network research. In publicity around this new paper, he has been careful to explain that the goal is to supplement human productivity rather than replace it.
READ the latest news shaping the AI Chips market at AI Chips News
AI Designed Computer Chips That the Human Mind Can’t Understand – Popular Mechanics, source
Add comment