Comments about "Artificial intelligence" in Wikipedia

This document contains comments about the document "Artificial intelligence" in Wikipedia
In the last paragraph Reflection I explain my own opinion.

Contents


Introduction

The article starts with the following sentence.
Artificial intelligence (AI) is the intelligence exhibited by machines. In computer science, an ideal "intelligent" machine is a flexible rational agent that perceives its environment and takes actions that maximize its chance of success at an arbitrary goal
The problem with this definition is that it includes terms which are not clear in this context. Typical words are: flexible, rational and perceive
Colloquially, the term "artificial intelligence" is likely to be applied when a machine uses cutting-edge techniques to competently perform or mimic "cognitive" functions that we intuitively associate with human minds, such as "learning" and "problem solving".
There are two issues:
  1. how does a PC learn? How does a PC solve a problem.
  2. is every problem that a PC can solve considered AI?
Modern examples of AI include computers that can beat professional players at Chess and Go, and self-driving cars that navigate crowded city streets.
It is all in the name.
What counts is to clasify chess playing programs accordingly to some criteria. When you do that you try evaluate the people behind these programs.

1. History

The field of AI research was founded at a conference on the campus of Dartmouth College in the summer of 1956. The attendees, including John McCarthy, Marvin Minsky, Allen Newell, Arthur Samuel, and Herbert Simon, became the leaders of AI research for many decades. They and their students wrote programs that were, to most people, simply astonishing: computers were winning at checkers, solving word problems in algebra, proving logical theorems and speaking English.
AI should be more than only writing a computer program, otherwise the concept of AI does not "make sense".
In the late 1990s and early 21st century, AI achieved its greatest successes, and began to be used for logistics, data mining, medical diagnosis and many other areas throughout the technology industry.
Again also in this sentence there is no indication what makes AI special. It seems that any automation project is an AI application.
A set of advanced statistical techniques (loosely known as deep learning), access to large amounts of data and faster computers produced enormous advances in machine learning and perception. By the mid 2010s, successful machine learning applications were being used widely throughout the world.
Faster and larger computer (networks) (should) have nothing to do with AI.
To link AI with machine learning, implying that the program adapts it self when it goes through the different execution cycles (iterations) is a concept that requires attention.

2 Research

2.1 Goals

2.2 Approaches

2.3 Tools

2.4 Evaluating progress

3. Applications

AI is relevant to any intellectual task
This sentence is a misnomer, irrelevant.

3.1 Competitions and prizes

4. Platforms

A platform (or "computing platform") is defined as "some sort of hardware architecture or software framework (including application frameworks), that allows software to run."
What is not mentioned is additional dedicated hardware that can be used to perform a certain task or operation. This is specific true in process control, however all what is used to control an airplane could also be included.
A software discussion does not belong in this section.

5. Philosophy and ethics

There are three philosophical questions related to AI:
1a. Is artificial general intelligence possible?
1b. Can a machine solve any problem that a human being can solve using intelligence?
1c. Or are there hard limits to what a machine can accomplish?
1a depents about the definition, but IMO does not make sense.
1b depents what you mean a. with a machine and b. to solve a problem.
A machine can be a PC, a grass-mower or robot which each can used in a very narow niche. To solve a problem is much more difficult because you need a clear description what is the problem and what is a solution.
Mathematical problems in general are much simpler as physical or economical problems. For many problems there are no solutions.
To solve mathematical problems with a computer you need an algorithm which clearly describes the calculations to perform in order to solve the problem.
1c is closely related to 1b. What is a machine?
If it is a PC than roughly speaking there are no hard limits.
If it is a Quantum Computer than may hard limits can be expected.
2a Are intelligent machines dangerous?
2b How can we ensure that machines behave ethically
2c and that they are used ethically?
2a An automatic rifle is dangerous. A process plant which manufactures poison can also be dangerous. A laboratory which does research on virusses can also be dangerous. An atomic power plant can also be dangerous.
2b is wrongly stated. Machine do not behave only humans behave.
2c is correct. The problem is what is meant with ethically.

5.1 The limits of artificial general intelligence

Can a machine be intelligent? Can it "think"?
A PC is not intelligent. It is the "programmer" who is intelligent. The PC can not think. It is the "programmer" thinks in order to develop the algorithm.

Turing's "polite convention"
We need not decide if a machine can "think"; we need only decide if a machine can act as intelligently as a human being.
The answer on this question depents how intelligent the interrogater is. All what he knows and his capability to reason i.e. to use his knowledge,
If the PC knows a specific time table and the interrogator not than the PC is more intelligent.
If the PC can not use the time table but the interrogator can than the interrogator is better.

The artificial brain argument
The brain can be simulated by machines and because brains are intelligent, simulated brains must also be intelligent; thus machines can be intelligent.
How do you know that a brain can be simulated by machines? This is a completely non proven assertion. Anyway how do you demonstrate that?

The AI effect
Machines are already intelligent, but observers have failed to recognize it
Such an adhoc statement does not make sense.
When Deep Blue beat Garry Kasparov in chess, the machine was acting intelligently. However, onlookers commonly discount the behavior of an artificial intelligence program by arguing that it is not "real" intelligence after all; thus "real" intelligence is whatever intelligent behavior people can do that machines still can not. This is known as the AI Effect: "AI is whatever hasn't been done yet."
When a computer program outperforms a human being it does not mean that the PC is intelligent. It means that the "programmer" who has written this program is more intelligent as the previous program. Ofcourse you should not compare apples with pears. See Reflection 1

5.2 Intelligent behaviour and machine ethics

As a minimum, an AI system must be able to reproduce aspects of human intelligence. This raises the issue of how ethically the machine should behave towards both humans and other AI agents.
The first sentence has "nothing" to do with the second.
Ethical behaviour is a human issue and not a machine or computer issue.
The issue is what type of devices can we humans build. The problem is many of the devices, tools, machines we build can be used in favour of our existance or in disfavour. This is clearly a human issue and not an AI issue.

5.3 Machine consciousness, sentience and mind

If an AI system replicates all key aspects of human intelligence, will that system also be sentient – will it have a mind which has conscious experiences? This question is closely related to the philosophical problem as to the nature of human consciousness, generally referred to as the hard problem of consciousness.
It is impossible to reproduce all human aspects in a machine including consiousness.
This what if question does not belong in this document.

5.4 Superintelligence

7. See also

Following is a list with "Comments in Wikipedia" about related subjects


Reflection

AI is more than just any computer program.
The problem with the concept of AI is that it is a relatif concept. The same with the concept intelligence which is strictly related to humans. Some humans are more intelligent than others. The issue is how do you that. How do you make better decisions based on the same information. The same with a computer program.
To evaluate which is the most intelligent program you should use certain rules.
  1. Computer memory is not an issue. That means the size of your program is not an issue.
  2. Computer speed is an issue. That means all programs should run on the same hardware.
  3. All programs should be based on the same data and solve the same applications. This is a tricky issue because programs should not contain existing knowledge except what is specific supplied as part of the application.
    For example: A chess playing program should not contain any opening sets if stated by the rules.

Feedback


If you want to give a comment you can use the following form Comment form
Created: 16 June 2016

Go Back to Wikipedia Comments in Wikipedia documents
Back to my home page Index