Neurons, astrocytes and transformers: Are AI models biologically plausible?

Advertisement · Scroll to continue

(ORDO NEWS) — Recent advances in neuroscience and machine learning have led to groundbreaking research that sheds light on the potential role of astrocytes in the human brain.

Astrocytes, a type of non-neuronal cell, have long remained a mystery to scientists, but new research from the Massachusetts Institute of Technology, the MIT-IBM Watson Artificial Intelligence Laboratory and Harvard Medical School offers a promising hypothesis that could bridge the gap between biology and artificial intelligence.

Transformers power

Transformers are a type of neural network models that have revolutionized the field of artificial intelligence. These models, modeled after the biological neurons that process information in the brain, have demonstrated unprecedented performance on tasks such as generating text with near-human accuracy.

However, the inner workings of Transformers have still not been revealed, making it difficult to understand how they could be created from biological components.

Unlocking the Potential of Astrocytes

Astrocytes are abundant in the brain and are known to interact with neurons and play a role in physiological processes. However, their computational function still remains unclear. In this paper, the researchers examined the computational role of astrocytes and proposed a mathematical model that suggests that astrocytes, along with neurons, can perform the same computations in the nucleus as a transformer.

The implications of this hypothesis are significant. Not only does it provide insight into how the human brain functions, but it also helps explain why Transformers are so successful at solving complex problems. By understanding the potential role of astrocytes, scientists will be able to further unlock the mysteries of the brain and advance the field of AI.

Neuroscience for AI and AI for Neuroscience

Dmitry Krotov, a postdoctoral fellow at the MIT-IBM Watson AI Lab and senior author of the paper, emphasizes the importance of bridging the gap between biology and AI.

He states: “The brain is far superior to even the best artificial neural networks we have developed, but we don’t know exactly how the brain works. The scientific value is in thinking about the connections between biological hardware and large-scale artificial intelligence networks. This is neuroscience for AI and AI for Neuroscience.”

The Future of Brain-Inspired Artificial Intelligence

This groundbreaking research opens new paths for future neuroscientific research and AI development. By harnessing the potential of astrocytes, scientists will be able to further understand the computational capabilities of the brain and potentially improve artificial intelligence systems.

As the study’s lead author Leo Kozachkov explained, “Our work suggests that astrocytes may be the missing piece of the puzzle in creating more biologically plausible AI systems.”

With this new idea, the field of artificial intelligence could take a major leap forward. By drawing inspiration from the complexities of the human brain, researchers will be able to develop more powerful and efficient AI systems that mimic the brain’s extraordinary capabilities.


Contact us: [email protected]

Our Standards, Terms of Use: Standard Terms And Conditions.

To eliminate any confusion arising from different time zones and daylight saving changes, all times displayed on our platforms are in Coordinated Universal Time (UTC).

Advertisement · Scroll to continue
Advertisement · Scroll to continue