China surpasses the US with a new AI model that is 100 times faster
Researchers at the Institute of Automation of the Chinese Academy of Sciences in Beijing developed the SpikingBrain 1.0 artificial intelligence language model, which mimics the functioning of the human brain and is 100 times faster than US models
A team of researchers at the Institute of Automation of the Chinese Academy of Sciences claims to have developed an artificial intelligence model that is 100 times faster than US models.
This is the SpikingBrain 1.0 model, which mimics the functioning of the brain and operates exclusively with local chips. In this regard, the scientists have emphasized that the model activates “only the neurons necessary for each task. This means that, instead of mobilizing its entire neural network like ChatGPT and other traditional AIs, it responds selectively to the stimulus received, which reduces energy consumption and significantly speeds up response times.”
Among the functions that SpikingBrain can perform, thanks to its biology-inspired design, is “learning from a very small amount of training data — less than 2% of what conventional systems require — and it continues to operate quickly and efficiently even when faced with very long texts.”
During testing, it achieved efficiency up to 100 times faster than traditional models. It is important to emphasize that the model’s advancement lies in its technological independence from the US, as it operates entirely on the “Chinese chip ecosystem, specifically on the MetaX platform, thus freeing it from the grip of Nvidia.”
Lead researcher Li Guoqi emphasized that the model “provides an optimized framework for Chinese chips” and that this AI “can be useful for processing long sequences of information such as legal documents, medical records, or scientific simulations. To democratize access, the team has released a smaller version and maintains a larger model available for public testing.”
One of the drawbacks of this model is the high energy consumption, not only for processing the large volume of data, but also for cooling the equipment.
M.Pino
Source: elconfidencial
(Reference image source: Growtika on Unsplash)
Visit our news channel on Google News and follow us to get accurate, interesting information and stay up to date with everything. You can also see our daily content on X/Twitter and Instagram
