Read a verse of “Spring Dawn” to the computer, and it can quickly continue to read it with “memory”; Spray insecticide at it, it can show the perceptual feedback of “pay attention to toxic gases”… On September 1, the reporter learned from Zhejiang University that Zhejiang University and Zhijiang Laboratory recently jointly developed China’s first brain-like computer (Darwin Mouse) based on independent intellectual property rights brain-like chips.
It is currently the largest neuronal brain-like computer in the world, containing 792 Darwin 2nd generation brain-like chips developed by Zhejiang University, supporting 1.200 million pulsed neurons, nearly 100 billion synapses, the number of neurons is equivalent to a mouse brain, greatly improving computing performance. Compared with the high energy consumption of traditional computers, the typical operating power consumption of this type of brain computer is only 350 watts to 500 watts.
At the same time, the team also developed the “housekeeper” – DarwinOS for brain-like computers, which is used to achieve effective management and scheduling of brain-like computer hardware resources and support the operation and application of brain-like computers.
The reporter saw at the experimental site that under the control of the intelligent center of the brain-like computer, three robots with similar appearance cooperated to carry out flood prevention and rescue task drills through voice movement instructions. Robot No. 1 patrolled the site with its own camera, and after finding a gap in the embankment, it called Robot No. 3 in charge of the project to repair the dam; After searching for a mannequin of a simulated wounded person who had fallen to the ground, Robot No. 1 called Robot No. 2, which was in charge of rescue, to come to the rescue.
“Using hardware and software to simulate the structure and operation mechanism of brain neural networks to construct a new artificial intelligence system, this new computing mode is brain-like computing.” It is characterized by integrated storage and computing, event-driven, and highly parallel. Pan Gang, the leader of the research team and a professor at the School of Computer Science and Technology of Zhejiang University, said that the transformation of computing mode and architecture is the biggest difference between brain-like computers and traditional computers. Brain-like computers are designed by drawing on biological brains, and algorithms are constantly changing, which makes brain-like computers expected to change the previous data-driven information processing methods, with lower power consumption and stronger intelligence, and are expected to complete tasks under more complex conditions.
In 2015 and 2019, Zhejiang University successfully developed Darwin 1 generation and Darwin 2 generation brain-like chips, respectively, and the result this time was to integrate hundreds of Darwin 2nd generation brain-like chips to form an “advanced” brain-like computer. At present, based on this kind of brain computer, researchers in the Hejiang Laboratory of Zhejiang University have realized a variety of intelligent tasks, including the collaborative work of multiple robots in flood prevention and rescue scenarios; A neural network model of the lateral geniculate nucleus of the thalamus was established by simulating a number of different brain regions by brain-like computer, simulating the periodic response of neurons in this brain region when flashing visual stimuli at different frequencies. The structure of neural circuits in the hippocampus and neural mechanisms were used to construct learning-memory fusion models.
Scientists said that in the future, brain-like computers can be used to process various intelligent tasks in life and open up application scenarios of artificial intelligence; It can also be used in neuroscience and brain science research, providing neuroscientists with faster and larger simulation tools and new experimental means to explore the working mechanism of the brain; It can also be implanted in mobile phones and robots to generate new intelligent service experiences.
(Source: Zhejiang Online – Zhejiang Daily)