From: Titanium Media
Human intelligence and learning ability are more advanced than machines than machines, and have surpassed geometry levels, and even reached the exponential level. How difficult is it for machines to cross from today's weak artificial intelligence to strong artificial intelligence? How many paradigms should be broken through? Can it be successful?
Author: Wang Ji, a cross-border high-yield author, an expert in over-the-counter finance and capital markets. We have paid attention to and researched the future trends of society and mankind for nearly 20 years. He has written more than 50 poems and has written 13 books on financial capital, natural science and management. The books he wrote have been ranked first in the "New Third Board" (Off-Counter Capital Market) category for no less than two and a half years, and 4 books have filled the gap in books in domestic capital-related industries. Publisher: Electronics Industry Publisher Publication time: April 2018
Differences between artificial intelligence and humans in learning ability
In order to objectively, vividly and deeply understand the concept of "the learning ability and intelligence of artificial intelligence are ten thousand times worse than that of humans or even billions of times", here is an analogy:
has two peach trees, namely A and B. Grafting different peaches on the branches of A can survive and basically grow peaches. However, if any other fruit is grafted, it will almost never survive, let alone grow the corresponding fruit.
If you want to get any fruit other than peaches, you either pick it directly from other trees (or vines), or plant it yourself, from seeds to seedlings to trees (vines), it is very troublesome to complete all the processes.
This is the real situation in reality. This is like the "intelligence" and "learning ability" of artificial intelligence, which are single or a little dimension. What a great miracle it would be if peach tree B, which grafted hundreds, thousands or even tens of thousands of different varieties of fruits (if there may be many branches) at the same time, if not only can they survive, but they can also grow almost tens of corresponding fruits (although the quality is uneven)! Is there such a situation in in nature? Surely not. However, this is the case with human intelligence and learning abilities.
Although a peach as big as a "basketball court" may grow on the "intellectual tree" of artificial intelligence (such as single-dimensional supercomputing power, chess calculation, text search ability, etc.), it can only grow peaches or a little similar fruit.
. Although you, as human, can only grow peaches as big as eggs (compared with single ability), your "intellectual power" can "graft" hundreds, thousands or even thousands of different fruits ( relies on the general intelligence of the brain to learn hundreds, thousands of skills , which is theoretically feasible). And these tens of thousands of fruits can also be linked into large networks, integrating, deriveing, and fissioning hundreds of millions of new fruits. What a great "universal intelligence" this is!
However, because everyone has this kind of intelligence, you seem to be very ordinary . Therefore, in daily life and study, people often despise it and even despise it. By comparing
, do you understand that human intelligence and learning abilities are hundreds, thousands or even tens of times stronger than machines?
In fact, this is not a big deal. There are two key points:
- Human beings can combine these hundreds and thousands of "intellectuals" to form higher, broader and deeper great "wisdom" across disciplines, fields and industries;
- Under the interaction of these wisdoms, human beings have achieved the greatest "collective learning" model in nature. How great and efficient is the model of
? If you look at the greatness of "human flesh search" and the vastness and greatness of knowledge sharing such as Wikipedia , you may feel a little taste - this is the most "unreasonable" learning ability!
Comparison of artificial intelligence and human intelligence levels
Here we briefly compare the differences between "intellectual intelligence", "intelligence" and "wisdom".
- Intellect
refers to the level of human cognitive ability and activity ability, which is mainly composed of observation ability, memory ability, thinking ability, imagination ability and operation ability.
- Intelligent
means having "intellect + ability". There are two meanings: one is "like having intelligence", which is much lower than intelligence, which is exactly the level of today's artificial intelligence; the other is "having the ability to use intelligence", which is much higher than intelligence, which is people's expectations for the future of artificial intelligence.
- Intelligence
is the highest level of things, that is, the ability to use intelligence to infinitely cross-border, deep learning, create ideas, innovate and invent (invent artificial intelligence), highly interactive, high-energy operations, efficient borrowing (such as using the super unilateral capabilities of artificial intelligence) and large-scale collaboration. The greatness of this ability is that it has unlimited imagination space, such as transforming the world and creating civilization.
In fact, in life and learning, a single individual does not need super low-level capabilities such as artificial intelligence (machines, systems and programs). If you need it, you can achieve it with the help of tools. Therefore, these capabilities that the machine possesses are the unique function of "tools". .
Obviously, it is precisely because of the ability of thousands and thousands of countless delicate branches of feelings and expressions, as well as the "wisdom" of thinking with countless characteristics, solving countless problems, etc., that we can adapt to and integrate into such a complex and diverse life in reality, create such a rich and complex human relationship, establish family, community, country and society, and create economy, politics, history, and such a profound and galaxy human civilization. If you compare
again, it may not be an exaggeration to say that human intelligence and learning ability are more advanced than machines than geometric series, and even reach the exponential level.
This also makes people understand how difficult it is for machines to cross from today's weak artificial intelligence to strong artificial intelligence (a wish), and how many paradigms should be broken through. Whether it can succeed is really a question that cannot be known.
In addition, for machine learning is not affected by human bias and has the advantage of less bias and never fatigue. Humans need to rest and sleep, and there are emotional problems such as jealousy, laziness, impulse and excitement. In fact, this is exactly the fundamental reason why machines are countless lower than humans (many people also regard it as a defect of humans, but in fact it is a misunderstanding of "not knowing honey in honey". They are so low that they don't know anything, and they only know how to do things in a single or a little dimension.
This also contrasts the complex "disadvantages" of human beings that only the "self" can have, and then have "thoughts" and "consciousness".
Let me say that, nowadays, artificial intelligence, which seems to be as powerful as myths and terrifying as myths, has not even had the intelligence and perception of 2 or 3-year-old children? The core reason is that these machines, systems or programs do not have the "consciousness" like humans at all.
In the future, with the continuous breakthroughs of technology, artificial intelligence will increasingly have the ability to "perception" in many aspects. However, these perceptions are just a "metaphor" of human beings, just like the "Chinese room experiment" proposed by the American philosopher John Searle (John Searle) in 1980 to refute the "Turing test".
We believe that only by breaking through "consciousness" can machines have the "universal" perception like humans, and then they can form ideas and produce universal strong artificial intelligence and super artificial intelligence like gods. Otherwise, artificial intelligence will be trapped in the category of "tools" like the swords, weapons and mobile phones in your hands.
So, can artificial intelligence break through the general paradigm of "consciousness"? Chapter 9 of the book "A Brief History of the New Future" uses more than 20,000 words to explain it to you in detail!
[Titanium Media Author Introduction: This article is excerpted from Wang Ji's book "A Brief History of the New Future".The author is a cross-border high-yield author and an expert in over-the-counter finance and capital markets. We have paid attention to and researched the future trends of society and mankind for nearly 20 years. He has written more than 50 poems and has written 13 books on financial capital, natural science and management. The books he wrote have been ranked first in JD's "New Third Board" (OTC Capital Market) book sales rankings for no less than two and a half years, and 4 books have filled the gaps in domestic capital-related industries. 】
