“Innovative components based on an oxide interface, developed by researchers at the ctd.qmat Cluster of Excellence in Würzburg, electronically replicate key functions of neural networks and open up new prospects for energy-efficient hardware. Credit: Image: Jochen Thamm, think-design. “(ScitechDaily, Scientists Create a Tiny Electronic Device That Can Learn Like the Human Brain)
AI changes math forever. What makes AI a tool. What can change the world of math forever is its ability to calculate axioms.
When the system tries to find zero points in things. Like. The Riemann conjecture (Zeta function). They. Need lots of calculations. They must find those zero points. In the lines of integers. From fractions between integers. There are infinitely many fractions between two integers. And that makes those calculations very hard. When. AI creates algorithms that should protect data. They can tell that. The algorithms should not use those zero points. What makes the AI a game changer is this. The same system can make many things.
But. As we know, AI requires new hardware. That hardware should be self-learning. And if we think about risks that we can see in those very small biomimicking computers, it is that. Those systems can connect multiple processors into one morphing neural network. The morphing means that the network. It can connect with multiple different devices. It can share itself. Into different. Separated groups. The thing that we should understand is that.
Those processors can control different types of drones and robots. When. Those processors start to operate as a network.
They can, surprisingly. Boost the system's data processing ability. There is possibility. That. AI that runs the system uses an “ask a friend” card. This means it can independently make contact with other devices. Those AIs can escape from their boxes. The situation is this. When those systems. Processor power is not enough. They can call for assistance. AI algorithms or agents can escape from their sandboxes. But in this case. That happens between physical devices.
If. We want to make a self-learning computer. The best way. It is to make a system. That mimics human brains.
That computer requires three layers. They will mimic the three main structures of the human brain. Those three layers are the forebrain, midbrain, and hindbrain. Those three main structures are the key. That makes our brains so effective. When a signal comes. From the hindbrain and RAS (reticular activating system) to the forebrain, it brings those signals into consciousness. The midbrain selects information that the brain stores in the molecular (DNA/RNA) based long-term memory. Another purpose of that central layer is to release or unpack the neural connections. If. Brains are in a jam. Or they are stuck in an eternal loop.
The three-layer system acts like the human brain. Researchers can make tiny systems that can mimic human brains. They can create three computer groups. There are things like supercomputers that replace those three layers. So tiny devices can form large structures. That means those three-layer devices. Or. Processors can form multiple internal three-layer groups. This means we could create a system that mimics human brains. By using three home computers. That means a very interesting future for that kind of system.
https://scitechdaily.com/scientists-create-a-tiny-electronic-device-that-can-learn-like-the-human-brain/
