Google says it has developed a way of using deep reinforcement learning (RL) to create computer chip floorplanning in just six hours — a complicated feat that typically requires humans months to achieve.
The chips Google's AI develops are on par or superior than those humans can create, the team explained in its paper published in the journal Nature on Wednesday, June 9.
In a first for one of its commercial products, Google's research is being used for the company's upcoming tensor processing unit (TPU) chips, which are optimized for AI computation.
So Google's AI method to design chips can eventually be used to improve and quicken the future development of AI.
"Our method was used to design the next generation of Google’s artificial intelligence (AI) accelerators, and has the potential to save thousands of hours of human effort for each new generation," the team said.
The major breakthrough is that Google's AI method can be used for chip "floorplanning," which, as the paper said "Despite five decades of research, chip floorplanning has defied automation, requiring months of intense effort by physical design engineers to produce manufacturable layouts."
Very nice work from Google on deep RL- based optimization for chip layout.— Yann LeCun (@ylecun) June 10, 2021
Simulated annealing and its heirs are finally dethroned after 40 years.
This uses graph NN and deConvNets, among other things.
I did not imagined back in the 90s that (de)ConvNets could be used for this. https://t.co/WY68QNnuMY
By using RL methods, which see an algorithm learn to execute certain actions so that it maximizes its chances of earning a reward, the Google team brought those months down to mere hours.
The team trained its AI system by feeding it 10,000 chip floorplans so that it could learn what was "right" and what was "wrong."
"As a result," the team said, "our method utilizes past experience to become better and faster at solving new instances of the problem, allowing chip design to be performed by artificial agents with more experience than any human designer."
The positive implications for semiconductor chips, in particular, the team notes, are high.
A lot of work has been happening on the semiconductor front lately. IBM, for instance, just had a major "semiconductor design breakthrough" by creating the world's first 2-nanometer chip. They're not the only ones having breakthroughs in tiny chip designs, as a team from MIT, the National Taiwan University, and the Taiwan Semiconductor Manufacturing Co. has found a way of manufacturing semiconductors below the 1-nm scale.