Dualism
Before we talk about creating consciousness we need to know a little bit about dualism, the idea that consciousness exists in some mental realm, separate from the physical realm. This belief is held by the majority of the world’s population and is held in Christianity, Islam and Buddhism. The conscious soul is sacred and cannot be duplicated.
But the opposite is true for general AI, which is materialistic. So consciousness can actually be created.
Whole brain simulation
Since the mind is thought to have consciousness, if the structure of a brain is replicated very precisely, it should also have consciousness. This is down to the level of fineness of each neuron and its connections.
The brain is arguably the most complex thing in the universe. The human brain has 85 billion neurons, and each neuron has thousands of connections between other neurons.
To replicate, the brain has to be scanned, and the resolution of images from medical scanners and MRI machines is only millimetres, which is too crude. To distinguish between brain cells at the nanoscale, only an electron microscope can be used.
Then there is the computing power of the human brain, which has to be equipped with exascale calculations because it can do them at speeds of up to one petaflop per second. At present, the supercomputing speed that human beings are working on is exascale, and the tianhe-3 E-class prototype that has just been accepted in China recently is at this level.
Finally, the model needs to be created based on all the brain cells that are scanned. This part is a bit of a hassle. We don’t know how to put the components together, how the modules will work together, etc. It’s not as simple as building a computer or a cell phone.
A previous project called “Open Worm” successfully created a worm brain that controlled a small robot and managed to get out of a Lego castle. But the worm’s brain is very simple, with just 302 neurons and 7,000 synapses.
Whole-brain simulations may not work because the brain is too complex. Maybe in addition to scanning the structure of all the brain cells, you also need to know the configuration of each cell, and maybe the layout of the axons and dendrites of each cell. In the worst case, you might want to know the position of every molecule or even atom. This is too complicated to implement. If there is a time dimension involved, the whole model is almost useless.
Based on narrow artificial intelligence
The narrow sense of ARTIFICIAL intelligence can roughly include symbolic artificial intelligence, machine learning, statistics and deep learning. The current stage of so-called artificial intelligence, also known as weak artificial intelligence.
Symbolic ARTIFICIAL intelligence, the beginning of artificial intelligence using symbols to represent the way of human thought, the typical representative is expert system.
Machine learning, which can derive results from data without writing a specific function manually. But people who use machine learning algorithms understand its limitations, and almost no one uses it to create conscious things. At most, it is to simulate and surpass human beings in a narrow field of human beings.
Statistics mainly refers to Bayesian networks, hidden Markov networks, Markov networks, etc., to model uncertain things.
Deep learning, in fact, is a kind of machine learning, because of the large number of layers of the network, so it is also called depth.
So the narrow ai approach says that a system can generate consciousness if it can perform enough machine learning operations. But the problem is that the human brain is completely different from other machine systems. A machine may be made up of different parts that work together. But the human brain is very different from many systems that share a single core.
Principle theory of consciousness
The idea is that we study the principles of human consciousness, understand the workings of the mind thoroughly, and then create a brain based on them. That would require humans to understand how the whole brain works before copying it, which obviously makes it very difficult.
Some people think that there is no need to understand the principle of the brain, because human beings used to imitate birds to fly, most of the principles are not understood, but human beings have been able to fly into space. So it’s not necessary to figure out how the whole brain works, and it may be possible to make artificial brains that are more powerful than the brain.
conclusion
The creation of consciousness depends on whether generalized artificial intelligence can be achieved, whether it can be generated by molecular replication of the entire brain, whether it can be generated by integrating thousands of machine learning systems, or whether it can be created by deciphering the principles of the entire brain.
Perhaps humans evolved consciousness slowly and inefficiently, but it would be much faster and more efficient to create consciousness scientifically.
————- Recommended reading ————
Summary of my open Source projects (Machine & Deep Learning, NLP, Network IO, AIML, mysql protocol, Chatbot)
Why to write “Analysis of Tomcat Kernel Design”
2018 summary data structure algorithms
2018 Summary machine learning
2018 Summary Java in Depth
2018 Summary of natural language processing
2018 Summary of deep learning
2018 summary JDK source code
2018 Summary Java concurrency Core
2018 summary reading passage
Talk to me, ask me questions:
Welcome to:
