In an office in Beijing at the Tsinghua University, a computer chip is crunching data that gets read from a nearby camera, looks for faces stored on a database and, seconds later, the exact same chip is handling voice commands in Chinese. The chip is known as Thinker, and it is designed to support neural networks. The special thing about Thinker is how little energy it uses, just eight AA batteries are enough to power it for a year.
Thinker can dynamically tailor its computing and memory requirements to meet the needs of the software being run. This is important since many real-world AI applications – recognising objects in images or understanding human speech – requires a combination of different kinds of neural networks with different numbers of layers.
This AI-chip is just the start of an important trend sweeping China's tech sector. It is a unique opportunity to establish itself amid the current wave of enthusiasm for hardware optimised for AI. Computer chips are key to the success of AI, so China needs to develop its own hardware industry to become a real force in the technology.
"Compared to how China responded to previous revolutions in information technology, the speed at which China is following the current [AI] trend is the fastest,” says Shouyi Yin, vice director of Tsinghua University’s Institute of Microelectronics and the lead author of the Thinker paper, referring to the effort to design neural-network processors in China.
It is possible to run AI software using existing chips such as the powerful graphics chips for FPGAs – a kind of blank chip that can be reconfigured on the fly – but those kind of designs are expensive and do not lend themselves to small devices that use batteries. That is initially why Yin's team at Tsinghua developed Thinker.
Thinker would be able to embed in a wide range of devices, such as smartphones, watches, home robots, or equipment stationed in remote areas. The team behind Thinker wants to launch the first product this coming March.
But Thinker is not the only project out there. Before February, a research team at the Chinese Academy of Sciences' Institute of Computing Technology (ICT) will have a local semiconductor manufacturer produce a small batch of chips for use in robots. This chip is called Dadu and has two cores – one for running neural networks and another for controlling motion. The neural core runs an algorithm for vision and allows the motion core to plan the optimal route for reaching a destination or the best motion for grabbing an object.
The director of the institute, Cyber Computing Lab, Yinhe Han, envisions a slew applications including robots that deliver coffee and drones controlled with hand gestures. The advantage of developing a system like this in China, Han says, is the large user base, which makes updating chip design based on user experience faster. China has tried and failed to update the chip industry before, and in 2001 the ICT assembled a team to develop desktop CPUs. That research team became the kernel of a Chinese chipmaker called Loongson, but the company's product never became as popular as the founders would have liked.
These Chinese chip startups find themselves in an environment that's vastly different from the one that gave birth to Intel or Nvidia. Businesses have taken to cloud computing in droves, basically, there may be less of a market for off-the-shelf hardware, says Dongrui Fan, president of SmartCo, a Beijing-based startup that designs an AI chip for data centres that process video footage. But China's AI companies are increasingly also developing their own hardware.
But for now, Chinese chip researchers have many problems to solve: how to commercialise their chip designs, how to scale up, and how to navigate a world of computing being transformed by AI. What’s not lacking, though, is the ambition for the future. "As chip researchers, we all have dreams," says Yinhe Han of ICT. "We’ll see how far we can leap."