Chapter 2281 Not That Empty

Ni Dongjun didn't expect Gu Weimian to say the names of his two teachers.

"What's the big deal? Where did I go to college? Who are my teachers? Want to find out? It's easy." He leaned against the head of the bed and softened, "Or do you just want to reminisce with me?"

Gu Weimian rolled his eyes out of nowhere, "Of course not, I'm not that free yet."

Ni Dongjun: "..."

He didn't expect the girl in front of him to be even crazier than him.

Is she begging him to chat now?

And she said she wasn't that free?

"If you really don't have time, then let's not talk for now.

Gu Weimian tilted her head, "The two main research directions of your current artificial intelligence industry are language and image, relying on data and integrated scale to achieve the main functions of speech recognition, graphic image recognition, etc."

She said in a low voice, not hurriedly, "And the foundation of all this is based on the existing computer neural network."

Ni Dongjun smiled slightly: "You know a lot."

He glanced at the time on his watch, "But, you just said a few words and it's been three minutes. "

The man reminded Gu Weiman that she didn't have time for what she said before.

Gu Weiman nodded, "Okay, it's just a pity that the irreversible disadvantage of computer neural networks compared to human biological neural networks can be solved.

Ni Dongjun was stunned, "Can it be solved?"

This is a bit too bold.

Artificial intelligence did not develop to its current level overnight.

There have been many times in the middle, scientists have tried and failed, and finally determined that the more optimized one is the current computer neural network.

With the rapid development of computer networks, there is also the development of computer hardware.

Once, many people thought that things that were impossible have now become a very popular and life-oriented thing.

These are all practical applications in life.

However, the existing computer neural networks have been constantly deified. Compared with the neurons that come with the human brain, they have an incomparable feature.

That is size.

To put it bluntly, computer neural networks are composed of transistors.

With the development of existing technology, these transistors are naturally getting smaller and smaller.

But compared with the neurons in the brain, it is far behind.

Although there are different opinions on the number of neurons in the human brain, the approximate order of magnitude is from billions to trillions.

Each neuron has an axon (output) and multiple dendrites (input), which means that it can output to multiple objects at the same time.

The basic working mode of brain neuron cells is through synaptic discharge and chemical reactions of the order of 100,000.

The biggest advantage of the transistor itself is calculation.

That is why the existing artificial intelligence needs to input a large amount of data to perform a task.

Let’s not talk about the process of data collection, just talk about the size issue.

If all the neurons in the human brain are replaced with transistors, then a building cannot be moved.

Compared with transistors, the disadvantage of the human brain is that it is slow.

But at the same time, compared with transistors, the human brain can manipulate situations that are hundreds of thousands or even billions of times more complex.

These things are of course very clear to Ni Dongjun, who has been immersed in this industry for more than ten years.